Merge branch 'access-pipeline' of gitlab.ub.uni-bielefeld.de:sfb1288inf/nopaque into access-pipeline

Dockerfile

@@ -4,11 +4,13 @@ FROM python:3.10.13-slim-bookworm
 LABEL authors="Patrick Jentsch <p.jentsch@uni-bielefeld.de>"
 
 
+# Set environment variables
 ENV LANG="C.UTF-8"
 ENV PYTHONDONTWRITEBYTECODE="1"
 ENV PYTHONUNBUFFERED="1"
 
 
+# Install system dependencies
 RUN apt-get update \
  && apt-get install --no-install-recommends --yes \
       build-essential \
@@ -17,37 +19,42 @@ RUN apt-get update \
  && rm --recursive /var/lib/apt/lists/*
 
 
+# Create a non-root user
 RUN useradd --create-home --no-log-init nopaque \
  && groupadd docker \
  && usermod --append --groups docker nopaque
 
-
 USER nopaque
 WORKDIR /home/nopaque
 
 
+# Create a Python virtual environment
 ENV NOPAQUE_PYTHON3_VENV_PATH="/home/nopaque/.venv"
 RUN python3 -m venv "${NOPAQUE_PYTHON3_VENV_PATH}"
 ENV PATH="${NOPAQUE_PYTHON3_VENV_PATH}/bin:${PATH}"
 
 
+# Install Python dependencies
+COPY --chown=nopaque:nopaque requirements.txt requirements.txt
+RUN python3 -m pip install --requirement requirements.txt \
+ && rm requirements.txt
+
+
+# Install the application
+COPY docker-nopaque-entrypoint.sh /usr/local/bin/
+
 COPY --chown=nopaque:nopaque app app
 COPY --chown=nopaque:nopaque migrations migrations
 COPY --chown=nopaque:nopaque tests tests
 COPY --chown=nopaque:nopaque .flaskenv boot.sh config.py nopaque.py requirements.txt ./
 
-
+RUN mkdir logs
-RUN python3 -m pip install --requirement requirements.txt \
- && mkdir logs
-
-
-USER root
-
-
-COPY docker-nopaque-entrypoint.sh /usr/local/bin/
 
 
 EXPOSE 5000
 
 
+USER root
+
+
 ENTRYPOINT ["docker-nopaque-entrypoint.sh"]

app/__init__.py

@@ -57,6 +57,9 @@ def create_app(config: Config = Config) -> Flask:
     scheduler.init_app(app)
     socketio.init_app(app, message_queue=app.config['NOPAQUE_SOCKETIO_MESSAGE_QUEUE_URI'])  # noqa
 
+    from .models.event_listeners import register_event_listeners
+    register_event_listeners()
+
     from .admin import bp as admin_blueprint
     default_breadcrumb_root(admin_blueprint, '.admin')
     app.register_blueprint(admin_blueprint, url_prefix='/admin')

app/contributions/spacy_nlp_pipeline_models/forms.py

@@ -16,8 +16,8 @@ class CreateSpaCyNLPPipelineModelForm(ContributionBaseForm):
     )
 
     def validate_spacy_model_file(self, field):
-        if not field.data.filename.lower().endswith('.tar.gz'):
+        if not field.data.filename.lower().endswith(('.tar.gz', ('.whl'))):
-            raise ValidationError('.tar.gz files only!')
+            raise ValidationError('.tar.gz or .whl files only!')
 
     def __init__(self, *args, **kwargs):
         if 'prefix' not in kwargs:

app/converters/sandpaper.py

@@ -2,80 +2,69 @@ from flask import current_app
 from app import db
 from app.models import User, Corpus, CorpusFile
 from datetime import datetime
+from pathlib import Path
+from typing import Dict, List
 import json
-import os
 import shutil
 
 
 class SandpaperConverter:
-    def __init__(self, json_db_file, data_dir):
+    def __init__(self, json_db_file: Path, data_dir: Path):
         self.json_db_file = json_db_file
         self.data_dir = data_dir
 
     def run(self):
-        with open(self.json_db_file, 'r') as f:
+        with self.json_db_file.open('r') as f:
-            json_db = json.loads(f.read())
+            json_db: List[Dict] = json.load(f)
 
         for json_user in json_db:
             if not json_user['confirmed']:
                 current_app.logger.info(f'Skip unconfirmed user {json_user["username"]}')
                 continue
-            user_dir = os.path.join(self.data_dir, str(json_user['id']))
+            user_dir = self.data_dir / f'{json_user["id"]}'
             self.convert_user(json_user, user_dir)
             db.session.commit()
 
 
-    def convert_user(self, json_user, user_dir):
+    def convert_user(self, json_user: Dict, user_dir: Path):
         current_app.logger.info(f'Create User {json_user["username"]}...')
-        user = User(
-            confirmed=json_user['confirmed'],
-            email=json_user['email'],
-            last_seen=datetime.fromtimestamp(json_user['last_seen']),
-            member_since=datetime.fromtimestamp(json_user['member_since']),
-            password_hash=json_user['password_hash'],  # TODO: Needs to be added manually
-            username=json_user['username']
-        )
-        db.session.add(user)
-        db.session.flush(objects=[user])
-        db.session.refresh(user)
         try:
-            user.makedirs()
+            user = User.create(
-        except OSError as e:
+                confirmed=json_user['confirmed'],
-            current_app.logger.error(e)
+                email=json_user['email'],
-            db.session.rollback()
+                last_seen=datetime.fromtimestamp(json_user['last_seen']),
+                member_since=datetime.fromtimestamp(json_user['member_since']),
+                password_hash=json_user['password_hash'],  # TODO: Needs to be added manually
+                username=json_user['username']
+            )
+        except OSError:
             raise Exception('Internal Server Error')
         for json_corpus in json_user['corpora'].values():
             if not json_corpus['files'].values():
                 current_app.logger.info(f'Skip empty corpus {json_corpus["title"]}')
                 continue
-            corpus_dir = os.path.join(user_dir, 'corpora', str(json_corpus['id']))
+            corpus_dir = user_dir / 'corpora' / f'{json_corpus["id"]}'
             self.convert_corpus(json_corpus, user, corpus_dir)
         current_app.logger.info('Done')
 
 
-    def convert_corpus(self, json_corpus, user, corpus_dir):
+    def convert_corpus(self, json_corpus: Dict, user: User, corpus_dir: Path):
         current_app.logger.info(f'Create Corpus {json_corpus["title"]}...')
-        corpus = Corpus(
-            user=user,
-            creation_date=datetime.fromtimestamp(json_corpus['creation_date']),
-            description=json_corpus['description'],
-            title=json_corpus['title']
-        )
-        db.session.add(corpus)
-        db.session.flush(objects=[corpus])
-        db.session.refresh(corpus)
         try:
-            corpus.makedirs()
+            corpus = Corpus.create(
-        except OSError as e:
+                user=user,
-            current_app.logger.error(e)
+                creation_date=datetime.fromtimestamp(json_corpus['creation_date']),
-            db.session.rollback()
+                description=json_corpus['description'],
+                title=json_corpus['title']
+            )
+        except OSError:
             raise Exception('Internal Server Error')
         for json_corpus_file in json_corpus['files'].values():
             self.convert_corpus_file(json_corpus_file, corpus, corpus_dir)
         current_app.logger.info('Done')
 
 
-    def convert_corpus_file(self, json_corpus_file, corpus, corpus_dir):
+    def convert_corpus_file(self, json_corpus_file: Dict, corpus: Corpus, corpus_dir: Path):
         current_app.logger.info(f'Create CorpusFile {json_corpus_file["title"]}...')
         corpus_file = CorpusFile(
             corpus=corpus,
@@ -99,13 +88,13 @@ class SandpaperConverter:
         db.session.refresh(corpus_file)
         try:
             shutil.copy2(
-                os.path.join(corpus_dir, json_corpus_file['filename']),
+                corpus_dir / json_corpus_file['filename'],
                 corpus_file.path
             )
         except:
             current_app.logger.warning(
                 'Can not convert corpus file: '
-                f'{os.path.join(corpus_dir, json_corpus_file["filename"])}'
+                f'{corpus_dir / json_corpus_file["filename"]}'
                 ' -> '
                 f'{corpus_file.path}'
             )

app/corpora/cli.py

@@ -1,7 +1,7 @@
-from app.models import Corpus, CorpusStatus
+from flask import current_app
-import os
 import shutil
 from app import db
+from app.models import Corpus, CorpusStatus
 from . import bp
 
 
@@ -18,10 +18,17 @@ def reset():
     ]
     for corpus in [x for x in Corpus.query.all() if x.status in status]:
         print(f'Resetting corpus {corpus}')
-        shutil.rmtree(os.path.join(corpus.path, 'cwb'), ignore_errors=True)
+        corpus_cwb_dir = corpus.path / 'cwb'
-        os.mkdir(os.path.join(corpus.path, 'cwb'))
+        corpus_cwb_data_dir = corpus_cwb_dir / 'data'
-        os.mkdir(os.path.join(corpus.path, 'cwb', 'data'))
+        corpus_cwb_registry_dir = corpus_cwb_dir / 'registry'
-        os.mkdir(os.path.join(corpus.path, 'cwb', 'registry'))
+        try:
+            shutil.rmtree(corpus.path / 'cwb', ignore_errors=True)
+            corpus_cwb_dir.mkdir()
+            corpus_cwb_data_dir.mkdir()
+            corpus_cwb_registry_dir.mkdir()
+        except OSError as e:
+            current_app.logger.error(e)
+            raise
         corpus.status = CorpusStatus.UNPREPARED
         corpus.num_analysis_sessions = 0
     db.session.commit()

app/corpora/cqi_over_sio/extensions.py

@@ -12,7 +12,6 @@ from typing import Dict, List
 import gzip
 import json
 import math
-import os
 from app import db
 from app.models import Corpus
 from .utils import lookups_by_cpos, partial_export_subcorpus, export_subcorpus
@@ -42,9 +41,9 @@ def ext_corpus_static_data(corpus: str) -> Dict:
     db_corpus_id: int = session['cqi_over_sio']['db_corpus_id']
     db_corpus: Corpus = Corpus.query.get(db_corpus_id)
 
-    static_data_file_path: str = os.path.join(db_corpus.path, 'cwb', 'static.json.gz')
+    static_data_file_path = db_corpus.path / 'cwb' / 'static.json.gz'
-    if os.path.exists(static_data_file_path):
+    if static_data_file_path.exists():
-        with open(static_data_file_path, 'rb') as f:
+        with static_data_file_path.open('rb') as f:
             return f.read()
 
     cqi_client: CQiClient = session['cqi_over_sio']['cqi_client']

app/corpora/files/routes.py

@@ -7,7 +7,6 @@ from flask import (
     url_for
 )
 from flask_breadcrumbs import register_breadcrumb
-import os
 from app import db
 from app.models import Corpus, CorpusFile, CorpusStatus
 from ..decorators import corpus_follower_permission_required
@@ -92,8 +91,8 @@ def corpus_file(corpus_id, corpus_file_id):
 def download_corpus_file(corpus_id, corpus_file_id):
     corpus_file = CorpusFile.query.filter_by(corpus_id=corpus_id, id=corpus_file_id).first_or_404()
     return send_from_directory(
-        os.path.dirname(corpus_file.path),
+        corpus_file.path.parent,
-        os.path.basename(corpus_file.path),
+        corpus_file.path.name,
         as_attachment=True,
         attachment_filename=corpus_file.filename,
         mimetype=corpus_file.mimetype

app/ext/flask_sqlalchemy/__init__.py

@@ -0,0 +1,2 @@
+from .container_column import ContainerColumn
+from .int_enum_column import IntEnumColumn

app/ext/flask_sqlalchemy/container_column.py

@@ -0,0 +1,21 @@
+import json
+from app import db
+
+
+class ContainerColumn(db.TypeDecorator):
+    impl = db.String
+
+    def __init__(self, container_type, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        self.container_type = container_type
+
+    def process_bind_param(self, value, dialect):
+        if isinstance(value, self.container_type):
+            return json.dumps(value)
+        elif isinstance(value, str) and isinstance(json.loads(value), self.container_type):
+            return value
+        else:
+            return TypeError()
+
+    def process_result_value(self, value, dialect):
+        return json.loads(value)

app/ext/flask_sqlalchemy/int_enum_column.py

@@ -0,0 +1,22 @@
+from app import db
+
+
+class IntEnumColumn(db.TypeDecorator):
+    impl = db.Integer
+
+    def __init__(self, enum_type, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        self.enum_type = enum_type
+
+    def process_bind_param(self, value, dialect):
+        if isinstance(value, self.enum_type) and isinstance(value.value, int):
+            return value.value
+        elif isinstance(value, int):
+            return self.enum_type(value).value
+        elif isinstance(value, str):
+            return self.enum_type[value].value
+        else:
+            return TypeError()
+
+    def process_result_value(self, value, dialect):
+        return self.enum_type(value)

app/jobs/json_routes.py

@@ -1,7 +1,6 @@
 from flask import abort, current_app
 from flask_login import current_user
 from threading import Thread
-import os
 from app import db
 from app.decorators import admin_required, content_negotiation
 from app.models import Job, JobStatus
@@ -39,7 +38,7 @@ def job_log(job_id):
     if job.status not in [JobStatus.COMPLETED, JobStatus.FAILED]:
         response = {'errors': {'message': 'Job status is not completed or failed'}}
         return response, 409
-    with open(os.path.join(job.path, 'pipeline_data', 'logs', 'pyflow_log.txt')) as log_file:
+    with open(job.path / 'pipeline_data' / 'logs' / 'pyflow_log.txt') as log_file:
         log = log_file.read()
     response_data = {
         'jobLog': log

app/jobs/routes.py

@@ -7,7 +7,6 @@ from flask import (
 )
 from flask_breadcrumbs import register_breadcrumb
 from flask_login import current_user
-import os
 from app.models import Job, JobInput, JobResult
 from . import bp
 from .utils import job_dynamic_list_constructor as job_dlc
@@ -38,8 +37,8 @@ def download_job_input(job_id, job_input_id):
     if not (job_input.job.user == current_user or current_user.is_administrator()):
         abort(403)
     return send_from_directory(
-        os.path.dirname(job_input.path),
+        job_input.path.parent,
-        os.path.basename(job_input.path),
+        job_input.path.name,
         as_attachment=True,
         attachment_filename=job_input.filename,
         mimetype=job_input.mimetype
@@ -52,8 +51,8 @@ def download_job_result(job_id, job_result_id):
     if not (job_result.job.user == current_user or current_user.is_administrator()):
         abort(403)
     return send_from_directory(
-        os.path.dirname(job_result.path),
+        job_result.path.parent,
-        os.path.basename(job_result.path),
+        job_result.path.name,
         as_attachment=True,
         attachment_filename=job_result.filename,
         mimetype=job_result.mimetype

app/main/cli.py

@@ -1,6 +1,7 @@
 from flask import current_app
 from flask_migrate import upgrade
-import os
+from pathlib import Path
+from typing import List
 from app.models import (
     CorpusFollowerRole,
     Role,
@@ -17,16 +18,15 @@ def deploy():
     # Make default directories
     print('Make default directories')
     base_dir = current_app.config['NOPAQUE_DATA_DIR']
-    default_dirs = [
+    default_dirs: List[Path] = [
-        os.path.join(base_dir, 'tmp'),
+        base_dir / 'tmp',
-        os.path.join(base_dir, 'users')
+        base_dir / 'users'
     ]
-    for dir in default_dirs:
+    for default_dir in default_dirs:
-        if os.path.exists(dir):
+        if not default_dir.exists():
-            if not os.path.isdir(dir):
+            default_dir.mkdir()
-                raise NotADirectoryError(f'{dir} is not a directory')
+        if not default_dir.is_dir():
-        else:
+            raise NotADirectoryError(f'{default_dir} is not a directory')
-            os.mkdir(dir)
 
     # migrate database to latest revision
     print('Migrate database to latest revision')

app/models/__init__.py

@@ -0,0 +1,19 @@
+from .avatar import *
+from .corpus_file import *
+from .corpus_follower_association import *
+from .corpus_follower_role import *
+from .corpus import *
+from .job_input import *
+from .job_result import *
+from .job import *
+from .role import *
+from .spacy_nlp_pipeline_model import *
+from .tesseract_ocr_pipeline_model import *
+from .token import *
+from .user import *
+from app import login
+
+
+@login.user_loader
+def load_user(user_id):
+    return User.query.get(int(user_id))

app/models/avatar.py

@@ -0,0 +1,40 @@
+from flask import current_app
+from flask_hashids import HashidMixin
+from pathlib import Path
+from app import db
+from .file_mixin import FileMixin
+
+
+class Avatar(HashidMixin, FileMixin, db.Model):
+    __tablename__ = 'avatars'
+    # Primary key
+    id = db.Column(db.Integer, primary_key=True)
+    # Foreign keys
+    user_id = db.Column(db.Integer, db.ForeignKey('users.id'))
+    # Relationships
+    user = db.relationship('User', back_populates='avatar')
+
+    @property
+    def path(self) -> Path:
+        return self.user.path / 'avatar'
+        # return os.path.join(self.user.path, 'avatar')
+
+    def delete(self):
+        try:
+            self.path.unlink(missing_ok=True)
+        except OSError as e:
+            current_app.logger.error(e)
+            raise
+        db.session.delete(self)
+
+    def to_json_serializeable(self, backrefs=False, relationships=False):
+        json_serializeable = {
+            'id': self.hashid,
+            **self.file_mixin_to_json_serializeable()
+        }
+        if backrefs:
+            json_serializeable['user'] = \
+                self.user.to_json_serializeable(backrefs=True)
+        if relationships:
+            pass
+        return json_serializeable

app/models/corpus.py

@@ -0,0 +1,200 @@
+from datetime import datetime
+from enum import IntEnum
+from flask import current_app, url_for
+from flask_hashids import HashidMixin
+from sqlalchemy.ext.associationproxy import association_proxy
+from typing import Union
+from pathlib import Path
+import shutil
+import xml.etree.ElementTree as ET
+from app import db
+from app.converters.vrt import normalize_vrt_file
+from app.ext.flask_sqlalchemy import IntEnumColumn
+from .corpus_follower_association import CorpusFollowerAssociation
+
+
+class CorpusStatus(IntEnum):
+    UNPREPARED = 1
+    SUBMITTED = 2
+    QUEUED = 3
+    BUILDING = 4
+    BUILT = 5
+    FAILED = 6
+    STARTING_ANALYSIS_SESSION = 7
+    RUNNING_ANALYSIS_SESSION = 8
+    CANCELING_ANALYSIS_SESSION = 9
+
+    @staticmethod
+    def get(corpus_status: Union['CorpusStatus', int, str]) -> 'CorpusStatus':
+        if isinstance(corpus_status, CorpusStatus):
+            return corpus_status
+        if isinstance(corpus_status, int):
+            return CorpusStatus(corpus_status)
+        if isinstance(corpus_status, str):
+            return CorpusStatus[corpus_status]
+        raise TypeError('corpus_status must be CorpusStatus, int, or str')
+
+
+class Corpus(HashidMixin, db.Model):
+    '''
+    Class to define a corpus.
+    '''
+    __tablename__ = 'corpora'
+    # Primary key
+    id = db.Column(db.Integer, primary_key=True)
+    # Foreign keys
+    user_id = db.Column(db.Integer, db.ForeignKey('users.id'))
+    # Fields
+    creation_date = db.Column(db.DateTime(), default=datetime.utcnow)
+    description = db.Column(db.String(255))
+    status = db.Column(
+        IntEnumColumn(CorpusStatus),
+        default=CorpusStatus.UNPREPARED
+    )
+    title = db.Column(db.String(32))
+    num_analysis_sessions = db.Column(db.Integer, default=0)
+    num_tokens = db.Column(db.Integer, default=0)
+    is_public = db.Column(db.Boolean, default=False)
+    # Relationships
+    files = db.relationship(
+        'CorpusFile',
+        back_populates='corpus',
+        lazy='dynamic',
+        cascade='all, delete-orphan'
+    )
+    corpus_follower_associations = db.relationship(
+        'CorpusFollowerAssociation',
+        back_populates='corpus',
+        cascade='all, delete-orphan'
+    )
+    followers = association_proxy(
+        'corpus_follower_associations',
+        'follower',
+        creator=lambda u: CorpusFollowerAssociation(follower=u)
+    )
+    user = db.relationship('User', back_populates='corpora')
+    # "static" attributes
+    max_num_tokens = 2_147_483_647
+
+    def __repr__(self):
+        return f'<Corpus {self.title}>'
+
+    @property
+    def analysis_url(self):
+        return url_for('corpora.analysis', corpus_id=self.id)
+
+    @property
+    def jsonpatch_path(self):
+        return f'{self.user.jsonpatch_path}/corpora/{self.hashid}'
+
+    @property
+    def path(self) -> Path:
+        return self.user.path / 'corpora' / f'{self.id}'
+
+    @property
+    def url(self):
+        return url_for('corpora.corpus', corpus_id=self.id)
+
+    @property
+    def user_hashid(self):
+        return self.user.hashid
+
+    @staticmethod
+    def create(**kwargs):
+        corpus = Corpus(**kwargs)
+        db.session.add(corpus)
+        db.session.flush(objects=[corpus])
+        db.session.refresh(corpus)
+        corpus_files_dir = corpus.path / 'files'
+        corpus_cwb_dir = corpus.path / 'cwb'
+        corpus_cwb_data_dir = corpus_cwb_dir / 'data'
+        corpus_cwb_registry_dir = corpus_cwb_dir / 'registry'
+        try:
+            corpus.path.mkdir()
+            corpus_files_dir.mkdir()
+            corpus_cwb_dir.mkdir()
+            corpus_cwb_data_dir.mkdir()
+            corpus_cwb_registry_dir.mkdir()
+        except OSError as e:
+            # TODO: Potential leftover cleanup
+            current_app.logger.error(e)
+            db.session.rollback()
+            raise
+        return corpus
+
+    def build(self):
+        corpus_cwb_dir = self.path / 'cwb'
+        corpus_cwb_data_dir = corpus_cwb_dir / 'data'
+        corpus_cwb_registry_dir = corpus_cwb_dir / 'registry'
+        try:
+            shutil.rmtree(corpus_cwb_dir, ignore_errors=True)
+            corpus_cwb_dir.mkdir()
+            corpus_cwb_data_dir.mkdir()
+            corpus_cwb_registry_dir.mkdir()
+        except OSError as e:
+            current_app.logger.error(e)
+            self.status = CorpusStatus.FAILED
+            raise
+        corpus_element = ET.fromstring('<corpus>\n</corpus>')
+        for corpus_file in self.files:
+            normalized_vrt_path = corpus_cwb_dir / f'{corpus_file.id}.norm.vrt'
+            try:
+                normalize_vrt_file(corpus_file.path, normalized_vrt_path)
+            except:
+                self.status = CorpusStatus.FAILED
+                return
+            element_tree = ET.parse(normalized_vrt_path)
+            text_element = element_tree.getroot()
+            text_element.set('author', corpus_file.author)
+            text_element.set('title', corpus_file.title)
+            text_element.set(
+                'publishing_year',
+                f'{corpus_file.publishing_year}'
+            )
+            text_element.set('address', corpus_file.address or 'NULL')
+            text_element.set('booktitle', corpus_file.booktitle or 'NULL')
+            text_element.set('chapter', corpus_file.chapter or 'NULL')
+            text_element.set('editor', corpus_file.editor or 'NULL')
+            text_element.set('institution', corpus_file.institution or 'NULL')
+            text_element.set('journal', corpus_file.journal or 'NULL')
+            text_element.set('pages', f'{corpus_file.pages}' or 'NULL')
+            text_element.set('publisher', corpus_file.publisher or 'NULL')
+            text_element.set('school', corpus_file.school or 'NULL')
+            text_element.tail = '\n'
+            # corpus_element.insert(1, text_element)
+            corpus_element.append(text_element)
+        ET.ElementTree(corpus_element).write(
+            corpus_cwb_dir / 'corpus.vrt',
+            encoding='utf-8'
+        )
+        self.status = CorpusStatus.SUBMITTED
+
+    def delete(self):
+        shutil.rmtree(self.path, ignore_errors=True)
+        db.session.delete(self)
+
+    def to_json_serializeable(self, backrefs=False, relationships=False):
+        json_serializeable = {
+            'id': self.hashid,
+            'creation_date': f'{self.creation_date.isoformat()}Z',
+            'description': self.description,
+            'max_num_tokens': self.max_num_tokens,
+            'num_analysis_sessions': self.num_analysis_sessions,
+            'num_tokens': self.num_tokens,
+            'status': self.status.name,
+            'title': self.title,
+            'is_public': self.is_public
+        }
+        if backrefs:
+            json_serializeable['user'] = \
+                self.user.to_json_serializeable(backrefs=True)
+        if relationships:
+            json_serializeable['corpus_follower_associations'] = {
+                x.hashid: x.to_json_serializeable()
+                for x in self.corpus_follower_associations
+            }
+            json_serializeable['files'] = {
+                x.hashid: x.to_json_serializeable(relationships=True)
+                for x in self.files
+            }
+        return json_serializeable

app/models/corpus_file.py

@@ -0,0 +1,102 @@
+from flask import current_app, url_for
+from flask_hashids import HashidMixin
+from pathlib import Path
+from app import db
+from .corpus import CorpusStatus
+from .file_mixin import FileMixin
+
+
+class CorpusFile(FileMixin, HashidMixin, db.Model):
+    __tablename__ = 'corpus_files'
+    # Primary key
+    id = db.Column(db.Integer, primary_key=True)
+    # Foreign keys
+    corpus_id = db.Column(db.Integer, db.ForeignKey('corpora.id'))
+    # Fields
+    author = db.Column(db.String(255))
+    description = db.Column(db.String(255))
+    publishing_year = db.Column(db.Integer)
+    title = db.Column(db.String(255))
+    address = db.Column(db.String(255))
+    booktitle = db.Column(db.String(255))
+    chapter = db.Column(db.String(255))
+    editor = db.Column(db.String(255))
+    institution = db.Column(db.String(255))
+    journal = db.Column(db.String(255))
+    pages = db.Column(db.String(255))
+    publisher = db.Column(db.String(255))
+    school = db.Column(db.String(255))
+    # Relationships
+    corpus = db.relationship(
+        'Corpus',
+        back_populates='files'
+    )
+
+    @property
+    def download_url(self):
+        return url_for(
+            'corpora.download_corpus_file',
+            corpus_id=self.corpus_id,
+            corpus_file_id=self.id
+        )
+
+    @property
+    def jsonpatch_path(self):
+        return f'{self.corpus.jsonpatch_path}/files/{self.hashid}'
+
+    @property
+    def path(self) -> Path:
+        return self.corpus.path / 'files' / f'{self.id}'
+
+    @property
+    def url(self):
+        return url_for(
+            'corpora.corpus_file',
+            corpus_id=self.corpus_id,
+            corpus_file_id=self.id
+        )
+
+    @property
+    def user_hashid(self):
+        return self.corpus.user.hashid
+
+    @property
+    def user_id(self):
+        return self.corpus.user_id
+
+    def delete(self):
+        try:
+            self.path.unlink(missing_ok=True)
+        except OSError as e:
+            current_app.logger.error(e)
+            raise
+        db.session.delete(self)
+        self.corpus.status = CorpusStatus.UNPREPARED
+
+    def to_json_serializeable(self, backrefs=False, relationships=False):
+        json_serializeable = {
+            'id': self.hashid,
+            'address': self.address,
+            'author': self.author,
+            'description': self.description,
+            'booktitle': self.booktitle,
+            'chapter': self.chapter,
+            'editor': self.editor,
+            'institution': self.institution,
+            'journal': self.journal,
+            'pages': self.pages,
+            'publisher': self.publisher,
+            'publishing_year': self.publishing_year,
+            'school': self.school,
+            'title': self.title,
+            **self.file_mixin_to_json_serializeable(
+                backrefs=backrefs,
+                relationships=relationships
+            )
+        }
+        if backrefs:
+            json_serializeable['corpus'] = \
+                self.corpus.to_json_serializeable(backrefs=True)
+        if relationships:
+            pass
+        return json_serializeable

app/models/corpus_follower_association.py

@@ -0,0 +1,47 @@
+from flask_hashids import HashidMixin
+from app import db
+from .corpus_follower_role import CorpusFollowerRole
+
+
+class CorpusFollowerAssociation(HashidMixin, db.Model):
+    __tablename__ = 'corpus_follower_associations'
+    # Primary key
+    id = db.Column(db.Integer, primary_key=True)
+    # Foreign keys
+    corpus_id = db.Column(db.Integer, db.ForeignKey('corpora.id'))
+    follower_id = db.Column(db.Integer, db.ForeignKey('users.id'))
+    role_id = db.Column(db.Integer, db.ForeignKey('corpus_follower_roles.id'))
+    # Relationships
+    corpus = db.relationship(
+        'Corpus',
+        back_populates='corpus_follower_associations'
+    )
+    follower = db.relationship(
+        'User',
+        back_populates='corpus_follower_associations'
+    )
+    role = db.relationship(
+        'CorpusFollowerRole',
+        back_populates='corpus_follower_associations'
+    )
+
+    def __init__(self, **kwargs):
+        if 'role' not in kwargs:
+            kwargs['role'] = CorpusFollowerRole.query.filter_by(default=True).first()
+        super().__init__(**kwargs)
+
+    def __repr__(self):
+        return f'<CorpusFollowerAssociation {self.follower.__repr__()} ~ {self.role.__repr__()} ~ {self.corpus.__repr__()}>'
+
+    def to_json_serializeable(self, backrefs=False, relationships=False):
+        json_serializeable = {
+            'id': self.hashid,
+            'corpus': self.corpus.to_json_serializeable(backrefs=True),
+            'follower': self.follower.to_json_serializeable(),
+            'role': self.role.to_json_serializeable()
+        }
+        if backrefs:
+            pass
+        if relationships:
+            pass
+        return json_serializeable

app/models/corpus_follower_role.py

@@ -0,0 +1,107 @@
+from flask_hashids import HashidMixin
+from enum import IntEnum
+from typing import Union
+from app import db
+
+
+class CorpusFollowerPermission(IntEnum):
+    VIEW = 1
+    MANAGE_FILES = 2
+    MANAGE_FOLLOWERS = 4
+    MANAGE_CORPUS = 8
+
+    @staticmethod
+    def get(corpus_follower_permission: Union['CorpusFollowerPermission', int, str]) -> 'CorpusFollowerPermission':
+        if isinstance(corpus_follower_permission, CorpusFollowerPermission):
+            return corpus_follower_permission
+        if isinstance(corpus_follower_permission, int):
+            return CorpusFollowerPermission(corpus_follower_permission)
+        if isinstance(corpus_follower_permission, str):
+            return CorpusFollowerPermission[corpus_follower_permission]
+        raise TypeError('corpus_follower_permission must be CorpusFollowerPermission, int, or str')
+
+
+class CorpusFollowerRole(HashidMixin, db.Model):
+    __tablename__ = 'corpus_follower_roles'
+    # Primary key
+    id = db.Column(db.Integer, primary_key=True)
+    # Fields
+    name = db.Column(db.String(64), unique=True)
+    default = db.Column(db.Boolean, default=False, index=True)
+    permissions = db.Column(db.Integer, default=0)
+    # Relationships
+    corpus_follower_associations = db.relationship(
+        'CorpusFollowerAssociation',
+        back_populates='role'
+    )
+
+    def __repr__(self):
+        return f'<CorpusFollowerRole {self.name}>'
+
+    def has_permission(self, permission: Union[CorpusFollowerPermission, int, str]):
+        perm = CorpusFollowerPermission.get(permission)
+        return self.permissions & perm.value == perm.value
+    
+    def add_permission(self, permission: Union[CorpusFollowerPermission, int, str]):
+        perm = CorpusFollowerPermission.get(permission)
+        if not self.has_permission(perm):
+            self.permissions += perm.value
+    
+    def remove_permission(self, permission: Union[CorpusFollowerPermission, int, str]):
+        perm = CorpusFollowerPermission.get(permission)
+        if self.has_permission(perm):
+            self.permissions -= perm.value
+
+    def reset_permissions(self):
+        self.permissions = 0
+
+    def to_json_serializeable(self, backrefs=False, relationships=False):
+        json_serializeable = {
+            'id': self.hashid,
+            'default': self.default,
+            'name': self.name,
+            'permissions': [
+                x.name
+                for x in CorpusFollowerPermission
+                if self.has_permission(x)
+            ]
+        }
+        if backrefs:
+            pass
+        if relationships:
+            json_serializeable['corpus_follower_association'] = {
+                x.hashid: x.to_json_serializeable(relationships=True)
+                for x in self.corpus_follower_association
+            }
+        return json_serializeable
+
+    @staticmethod
+    def insert_defaults():
+        roles = {
+            'Anonymous': [],
+            'Viewer': [
+                CorpusFollowerPermission.VIEW
+            ],
+            'Contributor': [
+                CorpusFollowerPermission.VIEW,
+                CorpusFollowerPermission.MANAGE_FILES
+            ],
+            'Administrator': [
+                CorpusFollowerPermission.VIEW,
+                CorpusFollowerPermission.MANAGE_FILES,
+                CorpusFollowerPermission.MANAGE_FOLLOWERS,
+                CorpusFollowerPermission.MANAGE_CORPUS
+
+            ]
+        }
+        default_role_name = 'Viewer'
+        for role_name, permissions in roles.items():
+            role = CorpusFollowerRole.query.filter_by(name=role_name).first()
+            if role is None:
+                role = CorpusFollowerRole(name=role_name)
+            role.reset_permissions()
+            for permission in permissions:
+                role.add_permission(permission)
+            role.default = role.name == default_role_name
+            db.session.add(role)
+        db.session.commit()

app/models/default_records/spacy_nlp_pipeline_model.yml

@@ -120,6 +120,7 @@
   version: '3.4.0'
   compatible_service_versions:
     - '0.1.1'
+    - '0.1.2'
 - title: 'German'
   description: 'German pipeline optimized for CPU. Components: tok2vec, tagger, morphologizer, parser, lemmatizer (trainable_lemmatizer), senter, ner.'
   url: 'https://github.com/explosion/spacy-models/releases/download/de_core_news_md-3.4.0/de_core_news_md-3.4.0.tar.gz'
@@ -131,6 +132,7 @@
   version: '3.4.0'
   compatible_service_versions:
     - '0.1.1'
+    - '0.1.2'
 - title: 'Greek'
   description: 'Greek pipeline optimized for CPU. Components: tok2vec, morphologizer, parser, lemmatizer (trainable_lemmatizer), senter, ner, attribute_ruler.'
   url: 'https://github.com/explosion/spacy-models/releases/download/el_core_news_md-3.4.0/el_core_news_md-3.4.0.tar.gz'
@@ -142,6 +144,7 @@
   version: '3.4.0'
   compatible_service_versions:
     - '0.1.1'
+    - '0.1.2'
 - title: 'English'
   description: 'English pipeline optimized for CPU. Components: tok2vec, tagger, parser, senter, ner, attribute_ruler, lemmatizer.'
   url: 'https://github.com/explosion/spacy-models/releases/download/en_core_web_md-3.4.1/en_core_web_md-3.4.1.tar.gz'
@@ -153,6 +156,7 @@
   version: '3.4.1'
   compatible_service_versions:
     - '0.1.1'
+    - '0.1.2'
 - title: 'Spanish'
   description: 'Spanish pipeline optimized for CPU. Components: tok2vec, morphologizer, parser, senter, ner, attribute_ruler, lemmatizer.'
   url: 'https://github.com/explosion/spacy-models/releases/download/es_core_news_md-3.4.0/es_core_news_md-3.4.0.tar.gz'
@@ -164,6 +168,7 @@
   version: '3.4.0'
   compatible_service_versions:
     - '0.1.1'
+    - '0.1.2'
 - title: 'French'
   description: 'French pipeline optimized for CPU. Components: tok2vec, morphologizer, parser, senter, ner, attribute_ruler, lemmatizer.'
   url: 'https://github.com/explosion/spacy-models/releases/download/fr_core_news_md-3.4.0/fr_core_news_md-3.4.0.tar.gz'
@@ -175,6 +180,7 @@
   version: '3.4.0'
   compatible_service_versions:
     - '0.1.1'
+    - '0.1.2'
 - title: 'Italian'
   description: 'Italian pipeline optimized for CPU. Components: tok2vec, morphologizer, tagger, parser, lemmatizer (trainable_lemmatizer), senter, ner'
   url: 'https://github.com/explosion/spacy-models/releases/download/it_core_news_md-3.4.0/it_core_news_md-3.4.0.tar.gz'
@@ -186,6 +192,7 @@
   version: '3.4.0'
   compatible_service_versions:
     - '0.1.1'
+    - '0.1.2'
 - title: 'Polish'
   description: 'Polish pipeline optimized for CPU. Components: tok2vec, morphologizer, parser, lemmatizer (trainable_lemmatizer), tagger, senter, ner.'
   url: 'https://github.com/explosion/spacy-models/releases/download/pl_core_news_md-3.4.0/pl_core_news_md-3.4.0.tar.gz'
@@ -197,6 +204,7 @@
   version: '3.4.0'
   compatible_service_versions:
     - '0.1.1'
+    - '0.1.2'
 - title: 'Russian'
   description: 'Russian pipeline optimized for CPU. Components: tok2vec, morphologizer, parser, senter, ner, attribute_ruler, lemmatizer.'
   url: 'https://github.com/explosion/spacy-models/releases/download/ru_core_news_md-3.4.0/ru_core_news_md-3.4.0.tar.gz'
@@ -208,6 +216,7 @@
   version: '3.4.0'
   compatible_service_versions:
     - '0.1.1'
+    - '0.1.2'
 - title: 'Chinese'
   description: 'Chinese pipeline optimized for CPU. Components: tok2vec, tagger, parser, senter, ner, attribute_ruler.'
   url: 'https://github.com/explosion/spacy-models/releases/download/zh_core_web_md-3.4.0/zh_core_web_md-3.4.0.tar.gz'
@@ -219,3 +228,4 @@
   version: '3.4.0'
   compatible_service_versions:
     - '0.1.1'
+    - '0.1.2'

app/models/event_listeners.py

@@ -0,0 +1,133 @@
+from datetime import datetime
+from enum import Enum
+from app import db, mail, socketio
+from app.email import create_message
+from .corpus_file import CorpusFile
+from .corpus_follower_association import CorpusFollowerAssociation
+from .corpus import Corpus
+from .job_input import JobInput
+from .job_result import JobResult
+from .job import Job, JobStatus
+from .spacy_nlp_pipeline_model import SpaCyNLPPipelineModel
+from .tesseract_ocr_pipeline_model import TesseractOCRPipelineModel
+from .user import UserSettingJobStatusMailNotificationLevel
+
+
+def register_event_listeners():
+    resources = [
+        Corpus,
+        CorpusFile,
+        Job,
+        JobInput,
+        JobResult,
+        SpaCyNLPPipelineModel,
+        TesseractOCRPipelineModel
+    ]
+
+    for resource in resources:
+        db.event.listen(resource, 'after_delete', resource_after_delete)
+        db.event.listen(resource, 'after_insert', resource_after_insert)
+        db.event.listen(resource, 'after_update', resource_after_update)
+
+    db.event.listen(CorpusFollowerAssociation, 'after_delete', cfa_after_delete)
+    db.event.listen(CorpusFollowerAssociation, 'after_insert', cfa_after_insert)
+
+    db.event.listen(Job, 'after_update', job_after_update)
+
+
+def resource_after_delete(mapper, connection, resource):
+    jsonpatch = [
+        {
+            'op': 'remove',
+            'path': resource.jsonpatch_path
+        }
+    ]
+    room = f'/users/{resource.user_hashid}'
+    socketio.emit('PATCH', jsonpatch, room=room)
+
+
+def cfa_after_delete(mapper, connection, cfa):
+    jsonpatch_path = f'/users/{cfa.corpus.user.hashid}/corpora/{cfa.corpus.hashid}/corpus_follower_associations/{cfa.hashid}'
+    jsonpatch = [
+        {
+            'op': 'remove',
+            'path': jsonpatch_path
+        }
+    ]
+    room = f'/users/{cfa.corpus.user.hashid}'
+    socketio.emit('PATCH', jsonpatch, room=room)
+
+
+def resource_after_insert(mapper, connection, resource):
+    jsonpatch_value = resource.to_json_serializeable()
+    for attr in mapper.relationships:
+        jsonpatch_value[attr.key] = {}
+    jsonpatch = [
+        {
+            'op': 'add',
+            'path': resource.jsonpatch_path,
+            'value': jsonpatch_value
+        }
+    ]
+    room = f'/users/{resource.user_hashid}'
+    socketio.emit('PATCH', jsonpatch, room=room)
+
+
+def cfa_after_insert(mapper, connection, cfa):
+    jsonpatch_value = cfa.to_json_serializeable()
+    jsonpatch_path = f'/users/{cfa.corpus.user.hashid}/corpora/{cfa.corpus.hashid}/corpus_follower_associations/{cfa.hashid}'
+    jsonpatch = [
+        {
+            'op': 'add',
+            'path': jsonpatch_path,
+            'value': jsonpatch_value
+        }
+    ]
+    room = f'/users/{cfa.corpus.user.hashid}'
+    socketio.emit('PATCH', jsonpatch, room=room)
+
+
+def resource_after_update(mapper, connection, resource):
+    jsonpatch = []
+    for attr in db.inspect(resource).attrs:
+        if attr.key in mapper.relationships:
+            continue
+        if not attr.load_history().has_changes():
+            continue
+        jsonpatch_path = f'{resource.jsonpatch_path}/{attr.key}'
+        if isinstance(attr.value, datetime):
+            jsonpatch_value = f'{attr.value.isoformat()}Z'
+        elif isinstance(attr.value, Enum):
+            jsonpatch_value = attr.value.name
+        else:
+            jsonpatch_value = attr.value
+        jsonpatch.append(
+            {
+                'op': 'replace',
+                'path': jsonpatch_path,
+                'value': jsonpatch_value
+            }
+        )
+    if jsonpatch:
+        room = f'/users/{resource.user_hashid}'
+        socketio.emit('PATCH', jsonpatch, room=room)
+
+
+def job_after_update(mapper, connection, job):
+    for attr in db.inspect(job).attrs:
+        if attr.key != 'status':
+            continue
+        if not attr.load_history().has_changes():
+            return
+        if job.user.setting_job_status_mail_notification_level == UserSettingJobStatusMailNotificationLevel.NONE:
+            return
+        if job.user.setting_job_status_mail_notification_level == UserSettingJobStatusMailNotificationLevel.END:
+            if job.status not in [JobStatus.COMPLETED, JobStatus.FAILED]:
+                return
+        msg = create_message(
+            job.user.email,
+            f'Status update for your Job "{job.title}"',
+            'tasks/email/notification',
+            job=job
+        )
+        mail.send(msg)

app/models/file_mixin.py

@@ -0,0 +1,40 @@
+from datetime import datetime
+from flask import current_app
+from werkzeug.utils import secure_filename
+from app import db
+
+
+class FileMixin:
+    '''
+    Mixin for db.Model classes. All file related models should use this.
+    '''
+    creation_date = db.Column(db.DateTime, default=datetime.utcnow)
+    filename = db.Column(db.String(255))
+    mimetype = db.Column(db.String(255))
+
+    def file_mixin_to_json_serializeable(self, backrefs=False, relationships=False):
+        return {
+            'creation_date': f'{self.creation_date.isoformat()}Z',
+            'filename': self.filename,
+            'mimetype': self.mimetype
+        }
+    
+    @classmethod
+    def create(cls, file_storage, **kwargs):
+        filename = kwargs.pop('filename', file_storage.filename)
+        mimetype = kwargs.pop('mimetype', file_storage.mimetype)
+        obj = cls(
+            filename=secure_filename(filename),
+            mimetype=mimetype,
+            **kwargs
+        )
+        db.session.add(obj)
+        db.session.flush(objects=[obj])
+        db.session.refresh(obj)
+        try:
+            file_storage.save(obj.path)
+        except (AttributeError, OSError) as e:
+            current_app.logger.error(e)
+            db.session.rollback()
+            raise e
+        return obj

app/models/job.py

@@ -0,0 +1,172 @@
+from datetime import datetime
+from enum import IntEnum
+from flask import current_app, url_for
+from flask_hashids import HashidMixin
+from time import sleep
+from typing import Union
+from pathlib import Path
+import shutil
+from app import db
+from app.ext.flask_sqlalchemy import ContainerColumn, IntEnumColumn
+
+
+class JobStatus(IntEnum):
+    INITIALIZING = 1
+    SUBMITTED = 2
+    QUEUED = 3
+    RUNNING = 4
+    CANCELING = 5
+    CANCELED = 6
+    COMPLETED = 7
+    FAILED = 8
+
+    @staticmethod
+    def get(job_status: Union['JobStatus', int, str]) -> 'JobStatus':
+        if isinstance(job_status, JobStatus):
+            return job_status
+        if isinstance(job_status, int):
+            return JobStatus(job_status)
+        if isinstance(job_status, str):
+            return JobStatus[job_status]
+        raise TypeError('job_status must be JobStatus, int, or str')
+
+
+class Job(HashidMixin, db.Model):
+    '''
+    Class to define Jobs.
+    '''
+    __tablename__ = 'jobs'
+    # Primary key
+    id = db.Column(db.Integer, primary_key=True)
+    # Foreign keys
+    user_id = db.Column(db.Integer, db.ForeignKey('users.id'))
+    # Fields
+    creation_date = \
+        db.Column(db.DateTime(), default=datetime.utcnow)
+    description = db.Column(db.String(255))
+    end_date = db.Column(db.DateTime())
+    service = db.Column(db.String(64))
+    service_args = db.Column(ContainerColumn(dict, 255))
+    service_version = db.Column(db.String(16))
+    status = db.Column(
+        IntEnumColumn(JobStatus),
+        default=JobStatus.INITIALIZING
+    )
+    title = db.Column(db.String(32))
+    # Relationships
+    inputs = db.relationship(
+        'JobInput',
+        back_populates='job',
+        cascade='all, delete-orphan',
+        lazy='dynamic'
+    )
+    results = db.relationship(
+        'JobResult',
+        back_populates='job',
+        cascade='all, delete-orphan',
+        lazy='dynamic'
+    )
+    user = db.relationship(
+        'User',
+        back_populates='jobs'
+    )
+
+    def __repr__(self):
+        return f'<Job {self.title}>'
+
+    @property
+    def jsonpatch_path(self):
+        return f'{self.user.jsonpatch_path}/jobs/{self.hashid}'
+
+    @property
+    def path(self) -> Path:
+        return self.user.path / 'jobs' / f'{self.id}'
+
+    @property
+    def url(self):
+        return url_for('jobs.job', job_id=self.id)
+
+    @property
+    def user_hashid(self):
+        return self.user.hashid
+
+    @staticmethod
+    def create(**kwargs):
+        job = Job(**kwargs)
+        db.session.add(job)
+        db.session.flush(objects=[job])
+        db.session.refresh(job)
+        job_inputs_dir = job.path / 'inputs'
+        job_pipeline_data_dir = job.path / 'pipeline_data'
+        job_results_dir = job.path / 'results'
+        try:
+            job.path.mkdir()
+            job_inputs_dir.mkdir()
+            job_pipeline_data_dir.mkdir()
+            job_results_dir.mkdir()
+        except OSError as e:
+            # TODO: Potential leftover cleanup
+            current_app.logger.error(e)
+            db.session.rollback()
+            raise
+        return job
+
+    def delete(self):
+        ''' Delete the job and its inputs and results from the database. '''
+        if self.status not in [JobStatus.COMPLETED, JobStatus.FAILED]:  # noqa
+            self.status = JobStatus.CANCELING
+            db.session.commit()
+            while self.status != JobStatus.CANCELED:
+                # In case the daemon handled a job in any way
+                if self.status != JobStatus.CANCELING:
+                    self.status = JobStatus.CANCELING
+                    db.session.commit()
+                sleep(1)
+                db.session.refresh(self)
+        try:
+            shutil.rmtree(self.path)
+        except OSError as e:
+            current_app.logger.error(e)
+            db.session.rollback()
+            raise e
+        db.session.delete(self)
+
+    def restart(self):
+        ''' Restart a job - only if the status is failed '''
+        if self.status != JobStatus.FAILED:
+            raise Exception('Job status is not "failed"')
+        shutil.rmtree(self.path / 'results', ignore_errors=True)
+        shutil.rmtree(self.path / 'pyflow.data', ignore_errors=True)
+        for result in self.results:
+            db.session.delete(result)
+        self.end_date = None
+        self.status = JobStatus.SUBMITTED
+
+    def to_json_serializeable(self, backrefs=False, relationships=False):
+        json_serializeable = {
+            'id': self.hashid,
+            'creation_date': f'{self.creation_date.isoformat()}Z',
+            'description': self.description,
+            'end_date': (
+                None if self.end_date is None
+                else f'{self.end_date.isoformat()}Z'
+            ),
+            'service': self.service,
+            'service_args': self.service_args,
+            'service_version': self.service_version,
+            'status': self.status.name,
+            'title': self.title
+        }
+        if backrefs:
+            json_serializeable['user'] = \
+                self.user.to_json_serializeable(backrefs=True)
+        if relationships:
+            json_serializeable['inputs'] = {
+                x.hashid: x.to_json_serializeable(relationships=True)
+                for x in self.inputs
+            }
+            json_serializeable['results'] = {
+                x.hashid: x.to_json_serializeable(relationships=True)
+                for x in self.results
+            }
+        return json_serializeable

app/models/job_input.py

@@ -0,0 +1,65 @@
+from flask import url_for
+from flask_hashids import HashidMixin
+from pathlib import Path
+from app import db
+from .file_mixin import FileMixin
+
+
+class JobInput(FileMixin, HashidMixin, db.Model):
+    __tablename__ = 'job_inputs'
+    # Primary key
+    id = db.Column(db.Integer, primary_key=True)
+    # Foreign keys
+    job_id = db.Column(db.Integer, db.ForeignKey('jobs.id'))
+    # Relationships
+    job = db.relationship(
+        'Job',
+        back_populates='inputs'
+    )
+
+    def __repr__(self):
+        return f'<JobInput {self.filename}>'
+
+    @property
+    def content_url(self):
+        return url_for(
+            'jobs.download_job_input',
+            job_id=self.job.id,
+            job_input_id=self.id
+        )
+
+    @property
+    def jsonpatch_path(self):
+        return f'{self.job.jsonpatch_path}/inputs/{self.hashid}'
+
+    @property
+    def path(self) -> Path:
+        return self.job.path / 'inputs' / f'{self.id}'
+
+    @property
+    def url(self):
+        return url_for(
+            'jobs.job',
+            job_id=self.job_id,
+            _anchor=f'job-{self.job.hashid}-input-{self.hashid}'
+        )
+
+    @property
+    def user_hashid(self):
+        return self.job.user.hashid
+
+    @property
+    def user_id(self):
+        return self.job.user.id
+
+    def to_json_serializeable(self, backrefs=False, relationships=False):
+        json_serializeable = {
+            'id': self.hashid,
+            **self.file_mixin_to_json_serializeable()
+        }
+        if backrefs:
+            json_serializeable['job'] = \
+                self.job.to_json_serializeable(backrefs=True)
+        if relationships:
+            pass
+        return json_serializeable

app/models/job_result.py

@@ -0,0 +1,71 @@
+from flask import url_for
+from flask_hashids import HashidMixin
+from pathlib import Path
+from app import db
+from .file_mixin import FileMixin
+
+
+class JobResult(FileMixin, HashidMixin, db.Model):
+    __tablename__ = 'job_results'
+    # Primary key
+    id = db.Column(db.Integer, primary_key=True)
+    # Foreign keys
+    job_id = db.Column(db.Integer, db.ForeignKey('jobs.id'))
+    # Fields
+    description = db.Column(db.String(255))
+    # Relationships
+    job = db.relationship(
+        'Job',
+        back_populates='results'
+    )
+
+    def __repr__(self):
+        return f'<JobResult {self.filename}>'
+
+    @property
+    def download_url(self):
+        return url_for(
+            'jobs.download_job_result',
+            job_id=self.job_id,
+            job_result_id=self.id
+        )
+
+    @property
+    def jsonpatch_path(self):
+        return f'{self.job.jsonpatch_path}/results/{self.hashid}'
+
+    @property
+    def path(self) -> Path:
+        return self.job.path / 'results' / f'{self.id}'
+
+    @property
+    def url(self):
+        return url_for(
+            'jobs.job',
+            job_id=self.job_id,
+            _anchor=f'job-{self.job.hashid}-result-{self.hashid}'
+        )
+
+    @property
+    def user_hashid(self):
+        return self.job.user.hashid
+
+    @property
+    def user_id(self):
+        return self.job.user.id
+
+    def to_json_serializeable(self, backrefs=False, relationships=False):
+        json_serializeable = {
+            'id': self.hashid,
+            'description': self.description,
+            **self.file_mixin_to_json_serializeable(
+                backrefs=backrefs,
+                relationships=relationships
+            )
+        }
+        if backrefs:
+            json_serializeable['job'] = \
+                self.job.to_json_serializeable(backrefs=True)
+        if relationships:
+            pass
+        return json_serializeable

app/models/role.py

@@ -0,0 +1,100 @@
+from enum import IntEnum
+from flask_hashids import HashidMixin
+from typing import Union
+from app import db
+
+
+class Permission(IntEnum):
+    '''
+    Defines User permissions as integers by the power of 2. User permission
+    can be evaluated using the bitwise operator &.
+    '''
+    ADMINISTRATE = 1
+    CONTRIBUTE = 2
+    USE_API = 4
+
+    @staticmethod
+    def get(permission: Union['Permission', int, str]) -> 'Permission':
+        if isinstance(permission, Permission):
+            return permission
+        if isinstance(permission, int):
+            return Permission(permission)
+        if isinstance(permission, str):
+            return Permission[permission]
+        raise TypeError('permission must be Permission, int, or str')
+
+
+class Role(HashidMixin, db.Model):
+    __tablename__ = 'roles'
+    # Primary key
+    id = db.Column(db.Integer, primary_key=True)
+    # Fields
+    name = db.Column(db.String(64), unique=True)
+    default = db.Column(db.Boolean, default=False, index=True)
+    permissions = db.Column(db.Integer, default=0)
+    # Relationships
+    users = db.relationship('User', back_populates='role', lazy='dynamic')
+
+    def __repr__(self):
+        return f'<Role {self.name}>'
+
+    def has_permission(self, permission: Union[Permission, int, str]):
+        p = Permission.get(permission)
+        return self.permissions & p.value == p.value
+    
+    def add_permission(self, permission: Union[Permission, int, str]):
+        p = Permission.get(permission)
+        if not self.has_permission(p):
+            self.permissions += p.value
+    
+    def remove_permission(self, permission: Union[Permission, int, str]):
+        p = Permission.get(permission)
+        if self.has_permission(p):
+            self.permissions -= p.value
+
+    def reset_permissions(self):
+        self.permissions = 0
+
+    def to_json_serializeable(self, backrefs=False, relationships=False):
+        json_serializeable = {
+            'id': self.hashid,
+            'default': self.default,
+            'name': self.name,
+            'permissions': [
+                x.name for x in Permission
+                if self.has_permission(x.value)
+            ]
+        }
+        if backrefs:
+            pass
+        if relationships:
+            json_serializeable['users'] = {
+                x.hashid: x.to_json_serializeable(relationships=True)
+                for x in self.users
+            }
+        return json_serializeable
+
+    @staticmethod
+    def insert_defaults():
+        roles = {
+            'User': [],
+            'API user': [Permission.USE_API],
+            'Contributor': [Permission.CONTRIBUTE],
+            'Administrator': [
+                Permission.ADMINISTRATE,
+                Permission.CONTRIBUTE,
+                Permission.USE_API
+            ],
+            'System user': []
+        }
+        default_role_name = 'User'
+        for role_name, permissions in roles.items():
+            role = Role.query.filter_by(name=role_name).first()
+            if role is None:
+                role = Role(name=role_name)
+            role.reset_permissions()
+            for permission in permissions:
+                role.add_permission(permission)
+            role.default = role.name == default_role_name
+            db.session.add(role)
+        db.session.commit()

app/models/spacy_nlp_pipeline_model.py

@@ -0,0 +1,136 @@
+from flask import current_app, url_for
+from flask_hashids import HashidMixin
+from tqdm import tqdm
+from pathlib import Path
+import requests
+import yaml
+from app import db
+from app.ext.flask_sqlalchemy import ContainerColumn
+from .file_mixin import FileMixin
+from .user import User
+
+
+class SpaCyNLPPipelineModel(FileMixin, HashidMixin, db.Model):
+    __tablename__ = 'spacy_nlp_pipeline_models'
+    # Primary key
+    id = db.Column(db.Integer, primary_key=True)
+    # Foreign keys
+    user_id = db.Column(db.Integer, db.ForeignKey('users.id'))
+    # Fields
+    title = db.Column(db.String(64))
+    description = db.Column(db.String(255))
+    version = db.Column(db.String(16))
+    compatible_service_versions = db.Column(ContainerColumn(list, 255))
+    publisher = db.Column(db.String(128))
+    publisher_url = db.Column(db.String(512))
+    publishing_url = db.Column(db.String(512))
+    publishing_year = db.Column(db.Integer)
+    pipeline_name = db.Column(db.String(64))
+    is_public = db.Column(db.Boolean, default=False)
+    # Relationships
+    user = db.relationship('User', back_populates='spacy_nlp_pipeline_models')
+
+    @property
+    def path(self) -> Path:
+        return self.user.path / 'spacy_nlp_pipeline_models' / f'{self.id}'
+
+    @property
+    def jsonpatch_path(self):
+        return f'{self.user.jsonpatch_path}/spacy_nlp_pipeline_models/{self.hashid}'
+
+    @property
+    def url(self):
+        return url_for(
+            'contributions.spacy_nlp_pipeline_model',
+            spacy_nlp_pipeline_model_id=self.id
+        )
+
+    @property
+    def user_hashid(self):
+        return self.user.hashid
+
+    @staticmethod
+    def insert_defaults(force_download=False):
+        nopaque_user = User.query.filter_by(username='nopaque').first()
+        default_records_file = Path(__file__).parent / 'default_records' / 'spacy_nlp_pipeline_model.yml'
+        with default_records_file.open('r') as f:
+            default_records = yaml.safe_load(f)
+        for m in default_records:
+            model = SpaCyNLPPipelineModel.query.filter_by(title=m['title'], version=m['version']).first()  # noqa
+            if model is not None:
+                model.compatible_service_versions = m['compatible_service_versions']
+                model.description = m['description']
+                model.filename = m['url'].split('/')[-1]
+                model.publisher = m['publisher']
+                model.publisher_url = m['publisher_url']
+                model.publishing_url = m['publishing_url']
+                model.publishing_year = m['publishing_year']
+                model.is_public = True
+                model.title = m['title']
+                model.version = m['version']
+                model.pipeline_name = m['pipeline_name']
+            else:
+                model = SpaCyNLPPipelineModel(
+                    compatible_service_versions=m['compatible_service_versions'],
+                    description=m['description'],
+                    filename=m['url'].split('/')[-1],
+                    publisher=m['publisher'],
+                    publisher_url=m['publisher_url'],
+                    publishing_url=m['publishing_url'],
+                    publishing_year=m['publishing_year'],
+                    is_public=True,
+                    title=m['title'],
+                    user=nopaque_user,
+                    version=m['version'],
+                    pipeline_name=m['pipeline_name']
+                )
+                db.session.add(model)
+                db.session.flush(objects=[model])
+                db.session.refresh(model)
+            if not model.path.exists() or force_download:
+                r = requests.get(m['url'], stream=True)
+                pbar = tqdm(
+                    desc=f'{model.title} ({model.filename})',
+                    unit="B",
+                    unit_scale=True,
+                    unit_divisor=1024,
+                    total=int(r.headers['Content-Length'])
+                )
+                pbar.clear()
+                with open(model.path, 'wb') as f:
+                    for chunk in r.iter_content(chunk_size=1024):
+                        if chunk:  # filter out keep-alive new chunks
+                            pbar.update(len(chunk))
+                            f.write(chunk)
+                    pbar.close()
+        db.session.commit()
+    
+    def delete(self):
+        try:
+            self.path.unlink(missing_ok=True)
+        except OSError as e:
+            current_app.logger.error(e)
+            raise
+        db.session.delete(self)
+
+    def to_json_serializeable(self, backrefs=False, relationships=False):
+        json_serializeable = {
+            'id': self.hashid,
+            'compatible_service_versions': self.compatible_service_versions,
+            'description': self.description,
+            'publisher': self.publisher,
+            'publisher_url': self.publisher_url,
+            'publishing_url': self.publishing_url,
+            'publishing_year': self.publishing_year,
+            'pipeline_name': self.pipeline_name,
+            'is_public': self.is_public,
+            'title': self.title,
+            'version': self.version,
+            **self.file_mixin_to_json_serializeable()
+        }
+        if backrefs:
+            json_serializeable['user'] = \
+                self.user.to_json_serializeable(backrefs=True)
+        if relationships:
+            pass
+        return json_serializeable

app/models/tesseract_ocr_pipeline_model.py

@@ -0,0 +1,132 @@
+from flask import current_app, url_for
+from flask_hashids import HashidMixin
+from tqdm import tqdm
+from pathlib import Path
+import requests
+import yaml
+from app import db
+from app.ext.flask_sqlalchemy import ContainerColumn
+from .file_mixin import FileMixin
+from .user import User
+
+
+class TesseractOCRPipelineModel(FileMixin, HashidMixin, db.Model):
+    __tablename__ = 'tesseract_ocr_pipeline_models'
+    # Primary key
+    id = db.Column(db.Integer, primary_key=True)
+    # Foreign keys
+    user_id = db.Column(db.Integer, db.ForeignKey('users.id'))
+    # Fields
+    title = db.Column(db.String(64))
+    description = db.Column(db.String(255))
+    version = db.Column(db.String(16))
+    compatible_service_versions = db.Column(ContainerColumn(list, 255))
+    publisher = db.Column(db.String(128))
+    publisher_url = db.Column(db.String(512))
+    publishing_url = db.Column(db.String(512))
+    publishing_year = db.Column(db.Integer)
+    is_public = db.Column(db.Boolean, default=False)
+    # Relationships
+    user = db.relationship('User', back_populates='tesseract_ocr_pipeline_models')
+
+    @property
+    def path(self) -> Path:
+        return self.user.path / 'tesseract_ocr_pipeline_models' / f'{self.id}'
+
+    @property
+    def jsonpatch_path(self):
+        return f'{self.user.jsonpatch_path}/tesseract_ocr_pipeline_models/{self.hashid}'
+
+    @property
+    def url(self):
+        return url_for(
+            'contributions.tesseract_ocr_pipeline_model',
+            tesseract_ocr_pipeline_model_id=self.id
+        )
+
+    @property
+    def user_hashid(self):
+        return self.user.hashid
+
+    @staticmethod
+    def insert_defaults(force_download=False):
+        nopaque_user = User.query.filter_by(username='nopaque').first()
+        default_records_file = Path(__file__).parent / 'default_records' / 'tesseract_ocr_pipeline_model.yml'
+        with default_records_file.open('r') as f:
+            default_records = yaml.safe_load(f)
+        for m in default_records:
+            model = TesseractOCRPipelineModel.query.filter_by(title=m['title'], version=m['version']).first()  # noqa
+            if model is not None:
+                model.compatible_service_versions = m['compatible_service_versions']
+                model.description = m['description']
+                model.filename = f'{model.id}.traineddata'
+                model.publisher = m['publisher']
+                model.publisher_url = m['publisher_url']
+                model.publishing_url = m['publishing_url']
+                model.publishing_year = m['publishing_year']
+                model.is_public = True
+                model.title = m['title']
+                model.version = m['version']
+            else:
+                model = TesseractOCRPipelineModel(
+                    compatible_service_versions=m['compatible_service_versions'],
+                    description=m['description'],
+                    publisher=m['publisher'],
+                    publisher_url=m['publisher_url'],
+                    publishing_url=m['publishing_url'],
+                    publishing_year=m['publishing_year'],
+                    is_public=True,
+                    title=m['title'],
+                    user=nopaque_user,
+                    version=m['version']
+                )
+                db.session.add(model)
+                db.session.flush(objects=[model])
+                db.session.refresh(model)
+                model.filename = f'{model.id}.traineddata'
+            if not model.path.exists() or force_download:
+                r = requests.get(m['url'], stream=True)
+                pbar = tqdm(
+                    desc=f'{model.title} ({model.filename})',
+                    unit="B",
+                    unit_scale=True,
+                    unit_divisor=1024,
+                    total=int(r.headers['Content-Length'])
+                )
+                pbar.clear()
+                with open(model.path, 'wb') as f:
+                    for chunk in r.iter_content(chunk_size=1024):
+                        if chunk:  # filter out keep-alive new chunks
+                            pbar.update(len(chunk))
+                            f.write(chunk)
+                    pbar.close()
+        db.session.commit()
+
+    def delete(self):
+        try:
+            self.path.unlink(missing_ok=True)
+        except OSError as e:
+            current_app.logger.error(e)
+            raise
+        db.session.delete(self)
+
+    def to_json_serializeable(self, backrefs=False, relationships=False):
+        json_serializeable = {
+            'id': self.hashid,
+            'compatible_service_versions': self.compatible_service_versions,
+            'description': self.description,
+            'publisher': self.publisher,
+            'publisher_url': self.publisher_url,
+            'publishing_url': self.publishing_url,
+            'publishing_year': self.publishing_year,
+            'is_public': self.is_public,
+            'title': self.title,
+            'version': self.version,
+            **self.file_mixin_to_json_serializeable()
+        }
+        if backrefs:
+            json_serializeable['user'] = \
+                self.user.to_json_serializeable(backrefs=True)
+        if relationships:
+            pass
+        return json_serializeable

app/models/token.py

@@ -0,0 +1,48 @@
+from datetime import datetime, timedelta
+from app import db
+
+
+class Token(db.Model):
+    __tablename__ = 'tokens'
+    # Primary key
+    id = db.Column(db.Integer, primary_key=True)
+    # Foreign keys
+    user_id = db.Column(db.Integer, db.ForeignKey('users.id'))
+    # Fields
+    access_token = db.Column(db.String(64), index=True)
+    access_expiration = db.Column(db.DateTime)
+    refresh_token = db.Column(db.String(64), index=True)
+    refresh_expiration = db.Column(db.DateTime)
+    # Relationships
+    user = db.relationship('User', back_populates='tokens')
+
+    def expire(self):
+        self.access_expiration = datetime.utcnow()
+        self.refresh_expiration = datetime.utcnow()
+
+    def to_json_serializeable(self, backrefs=False, relationships=False):
+        json_serializeable = {
+            'id': self.hashid,
+            'access_token': self.access_token,
+            'access_expiration': (
+                None if self.access_expiration is None
+                else f'{self.access_expiration.isoformat()}Z'
+            ),
+            'refresh_token': self.refresh_token,
+            'refresh_expiration': (
+                None if self.refresh_expiration is None
+                else f'{self.refresh_expiration.isoformat()}Z'
+            )
+        }
+        if backrefs:
+            json_serializeable['user'] = \
+                self.user.to_json_serializeable(backrefs=True)
+        if relationships:
+            pass
+        return json_serializeable
+
+    @staticmethod
+    def clean():
+        """Remove any tokens that have been expired for more than a day."""
+        yesterday = datetime.utcnow() - timedelta(days=1)
+        Token.query.filter(Token.refresh_expiration < yesterday).delete()

app/models/user.py

@@ -0,0 +1,452 @@
+from datetime import datetime, timedelta
+from enum import IntEnum
+from flask import current_app, url_for
+from flask_hashids import HashidMixin
+from flask_login import UserMixin
+from sqlalchemy.ext.associationproxy import association_proxy
+from pathlib import Path
+from typing import Union
+from werkzeug.security import generate_password_hash, check_password_hash
+import jwt
+import re
+import secrets
+import shutil
+from app import db, hashids
+from app.ext.flask_sqlalchemy import IntEnumColumn
+from .corpus import Corpus
+from .corpus_follower_association import CorpusFollowerAssociation
+from .corpus_follower_role import CorpusFollowerRole
+from .role import Permission, Role
+from .token import Token
+
+
+class ProfilePrivacySettings(IntEnum):
+    SHOW_EMAIL = 1
+    SHOW_LAST_SEEN = 2
+    SHOW_MEMBER_SINCE = 4
+
+    @staticmethod
+    def get(profile_privacy_setting: Union['ProfilePrivacySettings', int, str]) -> 'ProfilePrivacySettings':
+        if isinstance(profile_privacy_setting, ProfilePrivacySettings):
+            return profile_privacy_setting
+        if isinstance(profile_privacy_setting, int):
+            return ProfilePrivacySettings(profile_privacy_setting)
+        if isinstance(profile_privacy_setting, str):
+            return ProfilePrivacySettings[profile_privacy_setting]
+        raise TypeError('profile_privacy_setting must be ProfilePrivacySettings, int, or str')
+
+
+class UserSettingJobStatusMailNotificationLevel(IntEnum):
+    NONE = 1
+    END = 2
+    ALL = 3
+
+
+class User(HashidMixin, UserMixin, db.Model):
+    __tablename__ = 'users'
+    # Primary key
+    id = db.Column(db.Integer, primary_key=True)
+    # Foreign keys
+    role_id = db.Column(db.Integer, db.ForeignKey('roles.id'))
+    # Fields
+    email = db.Column(db.String(254), index=True, unique=True)
+    username = db.Column(db.String(64), index=True, unique=True)
+    username_pattern = re.compile(r'^[A-Za-zÄÖÜäöüß0-9_.]*$')
+    password_hash = db.Column(db.String(128))
+    confirmed = db.Column(db.Boolean, default=False)
+    terms_of_use_accepted = db.Column(db.Boolean, default=False)
+    member_since = db.Column(db.DateTime(), default=datetime.utcnow)
+    setting_job_status_mail_notification_level = db.Column(
+        IntEnumColumn(UserSettingJobStatusMailNotificationLevel),
+        default=UserSettingJobStatusMailNotificationLevel.END
+    )
+    last_seen = db.Column(db.DateTime())
+    full_name = db.Column(db.String(64))
+    about_me = db.Column(db.String(256))
+    location = db.Column(db.String(64))
+    website = db.Column(db.String(128))
+    organization = db.Column(db.String(128))
+    is_public = db.Column(db.Boolean, default=False)
+    profile_privacy_settings = db.Column(db.Integer(), default=0)
+    # Relationships
+    avatar = db.relationship(
+        'Avatar',
+        back_populates='user',
+        cascade='all, delete-orphan',
+        uselist=False
+    )
+    corpora = db.relationship(
+        'Corpus',
+        back_populates='user',
+        cascade='all, delete-orphan',
+        lazy='dynamic'
+    )
+    corpus_follower_associations = db.relationship(
+        'CorpusFollowerAssociation',
+        back_populates='follower',
+        cascade='all, delete-orphan'
+    )
+    followed_corpora = association_proxy(
+        'corpus_follower_associations',
+        'corpus',
+        creator=lambda c: CorpusFollowerAssociation(corpus=c)
+    )
+    jobs = db.relationship(
+        'Job',
+        back_populates='user',
+        cascade='all, delete-orphan',
+        lazy='dynamic'
+    )
+    role = db.relationship(
+        'Role',
+        back_populates='users'
+    )
+    spacy_nlp_pipeline_models = db.relationship(
+        'SpaCyNLPPipelineModel',
+        back_populates='user',
+        cascade='all, delete-orphan',
+        lazy='dynamic'
+    )
+    tesseract_ocr_pipeline_models = db.relationship(
+        'TesseractOCRPipelineModel',
+        back_populates='user',
+        cascade='all, delete-orphan',
+        lazy='dynamic'
+    )
+    tokens = db.relationship(
+        'Token',
+        back_populates='user',
+        cascade='all, delete-orphan',
+        lazy='dynamic'
+    )
+
+    def __init__(self, **kwargs):
+        if 'role' not in kwargs:
+            kwargs['role'] = (
+                Role.query.filter_by(name='Administrator').first()
+                if kwargs['email'] == current_app.config['NOPAQUE_ADMIN']
+                else Role.query.filter_by(default=True).first()
+            )
+        super().__init__(**kwargs)
+
+    def __repr__(self):
+        return f'<User {self.username}>'
+
+    @property
+    def jsonpatch_path(self):
+        return f'/users/{self.hashid}'
+
+    @property
+    def password(self):
+        raise AttributeError('password is not a readable attribute')
+
+    @password.setter
+    def password(self, password):
+        self.password_hash = generate_password_hash(password)
+
+    @property
+    def path(self) -> Path:
+        return current_app.config.get('NOPAQUE_DATA_DIR') / 'users' / f'{self.id}'
+
+    @staticmethod
+    def create(**kwargs):
+        user = User(**kwargs)
+        db.session.add(user)
+        db.session.flush(objects=[user])
+        db.session.refresh(user)
+        user_spacy_nlp_pipeline_models_dir = user.path / 'spacy_nlp_pipeline_models'
+        user_tesseract_ocr_pipeline_models_dir = user.path / 'tesseract_ocr_pipeline_models'
+        user_corpora_dir = user.path / 'corpora'
+        user_jobs_dir = user.path / 'jobs'
+        try:
+            user.path.mkdir()
+            user_spacy_nlp_pipeline_models_dir.mkdir()
+            user_tesseract_ocr_pipeline_models_dir.mkdir()
+            user_corpora_dir.mkdir()
+            user_jobs_dir.mkdir()
+        except OSError as e:
+            # TODO: Potential leftover cleanup
+            current_app.logger.error(e)
+            db.session.rollback()
+            raise
+        return user
+
+    @staticmethod
+    def insert_defaults():
+        nopaque_user = User.query.filter_by(username='nopaque').first()
+        system_user_role = Role.query.filter_by(name='System user').first()
+        if nopaque_user is None:
+            nopaque_user = User.create(
+                username='nopaque',
+                role=system_user_role
+            )
+            db.session.add(nopaque_user)
+        elif nopaque_user.role != system_user_role:
+            nopaque_user.role = system_user_role
+        db.session.commit()
+
+    @staticmethod
+    def reset_password(token, new_password):
+        try:
+            payload = jwt.decode(
+                token,
+                current_app.config['SECRET_KEY'],
+                algorithms=['HS256'],
+                issuer=current_app.config['SERVER_NAME'],
+                options={'require': ['exp', 'iat', 'iss', 'purpose', 'sub']}
+            )
+        except jwt.PyJWTError:
+            return False
+        if payload.get('purpose') != 'User.reset_password':
+            return False
+        user_hashid = payload.get('sub')
+        user_id = hashids.decode(user_hashid)
+        user = User.query.get(user_id)
+        if user is None:
+            return False
+        user.password = new_password
+        db.session.add(user)
+        return True
+
+    @staticmethod
+    def verify_access_token(access_token, refresh_token=None):
+        token = Token.query.filter(Token.access_token == access_token).first()
+        if token is not None:
+            if token.access_expiration > datetime.utcnow():
+                token.user.ping()
+                db.session.commit()
+                if token.user.role.name != 'System user':
+                    return token.user
+
+    @staticmethod
+    def verify_refresh_token(refresh_token, access_token):
+        token = Token.query.filter((Token.refresh_token == refresh_token) & (Token.access_token == access_token)).first()
+        if token is not None:
+            if token.refresh_expiration > datetime.utcnow():
+                return token
+            # someone tried to refresh with an expired token
+            # revoke all tokens from this user as a precaution
+            token.user.revoke_auth_tokens()
+            db.session.commit()
+
+    def can(self, permission):
+        return self.role is not None and self.role.has_permission(permission)
+
+    def confirm(self, confirmation_token):
+        try:
+            payload = jwt.decode(
+                confirmation_token,
+                current_app.config['SECRET_KEY'],
+                algorithms=['HS256'],
+                issuer=current_app.config['SERVER_NAME'],
+                options={'require': ['exp', 'iat', 'iss', 'purpose', 'sub']}
+            )
+        except jwt.PyJWTError:
+            return False
+        if payload.get('purpose') != 'user.confirm':
+            return False
+        if payload.get('sub') != self.hashid:
+            return False
+        self.confirmed = True
+        db.session.add(self)
+        return True
+
+    def delete(self):
+        shutil.rmtree(self.path, ignore_errors=True)
+        db.session.delete(self)
+
+    def generate_auth_token(self):
+        return Token(
+            access_token=secrets.token_urlsafe(),
+            access_expiration=datetime.utcnow() + timedelta(minutes=15),
+            refresh_token=secrets.token_urlsafe(),
+            refresh_expiration=datetime.utcnow() + timedelta(days=7),
+            user=self
+        )
+
+    def generate_confirm_token(self, expiration=3600):
+        now = datetime.utcnow()
+        payload = {
+            'exp': now + timedelta(seconds=expiration),
+            'iat': now,
+            'iss': current_app.config['SERVER_NAME'],
+            'purpose': 'user.confirm',
+            'sub': self.hashid
+        }
+        return jwt.encode(
+            payload,
+            current_app.config['SECRET_KEY'],
+            algorithm='HS256'
+        )
+
+    def generate_reset_password_token(self, expiration=3600):
+        now = datetime.utcnow()
+        payload = {
+            'exp': now + timedelta(seconds=expiration),
+            'iat': now,
+            'iss': current_app.config['SERVER_NAME'],
+            'purpose': 'User.reset_password',
+            'sub': self.hashid
+        }
+        return jwt.encode(
+            payload,
+            current_app.config['SECRET_KEY'],
+            algorithm='HS256'
+        )
+
+    def is_administrator(self):
+        return self.can(Permission.ADMINISTRATE)
+
+    def ping(self):
+        self.last_seen = datetime.utcnow()
+
+    def revoke_auth_tokens(self):
+        for token in self.tokens:
+            db.session.delete(token)
+
+    def verify_password(self, password):
+        if self.role.name == 'System user':
+            return False
+        return check_password_hash(self.password_hash, password)
+
+    #region Profile Privacy settings
+    def has_profile_privacy_setting(self, setting):
+        s = ProfilePrivacySettings.get(setting)
+        return self.profile_privacy_settings & s.value == s.value
+    
+    def add_profile_privacy_setting(self, setting):
+        s = ProfilePrivacySettings.get(setting)
+        if not self.has_profile_privacy_setting(s):
+            self.profile_privacy_settings += s.value
+
+    def remove_profile_privacy_setting(self, setting):
+        s = ProfilePrivacySettings.get(setting)
+        if self.has_profile_privacy_setting(s):
+            self.profile_privacy_settings -= s.value
+
+    def reset_profile_privacy_settings(self):
+        self.profile_privacy_settings = 0
+    #endregion Profile Privacy settings
+
+    def follow_corpus(self, corpus, role=None):
+        if role is None:
+            cfr = CorpusFollowerRole.query.filter_by(default=True).first()
+        else:
+            cfr = role
+        if self.is_following_corpus(corpus):
+            cfa = CorpusFollowerAssociation.query.filter_by(corpus=corpus, follower=self).first()
+            if cfa.role != cfr:
+                cfa.role = cfr
+        else:
+            cfa = CorpusFollowerAssociation(corpus=corpus, role=cfr, follower=self)
+            db.session.add(cfa)
+
+    def unfollow_corpus(self, corpus):
+        if not self.is_following_corpus(corpus):
+            return
+        self.followed_corpora.remove(corpus)
+
+    def is_following_corpus(self, corpus):
+        return corpus in self.followed_corpora
+    
+    def generate_follow_corpus_token(self, corpus_hashid, role_name, expiration=7):
+        now = datetime.utcnow()
+        payload = {
+            'exp': expiration,
+            'iat': now,
+            'iss': current_app.config['SERVER_NAME'],
+            'purpose': 'User.follow_corpus',
+            'role_name': role_name,
+            'sub': corpus_hashid
+        }
+        return jwt.encode(
+            payload,
+            current_app.config['SECRET_KEY'],
+            algorithm='HS256'
+        )
+    
+    def follow_corpus_by_token(self, token):
+        try:
+            payload = jwt.decode(
+                token,
+                current_app.config['SECRET_KEY'],
+                algorithms=['HS256'],
+                issuer=current_app.config['SERVER_NAME'],
+                options={'require': ['exp', 'iat', 'iss', 'purpose', 'role_name', 'sub']}
+            )
+        except jwt.PyJWTError:
+            return False
+        if payload.get('purpose') != 'User.follow_corpus':
+            return False
+        corpus_hashid = payload.get('sub')
+        corpus_id = hashids.decode(corpus_hashid)
+        corpus = Corpus.query.get_or_404(corpus_id)
+        if corpus is None:
+            return False
+        role_name = payload.get('role_name')
+        role = CorpusFollowerRole.query.filter_by(name=role_name).first()
+        if role is None:
+            return False
+        self.follow_corpus(corpus, role)
+        # db.session.add(self)
+        return True
+
+    def to_json_serializeable(self, backrefs=False, relationships=False, filter_by_privacy_settings=False):
+        json_serializeable = {
+            'id': self.hashid,
+            'confirmed': self.confirmed,
+            'avatar': url_for('users.user_avatar', user_id=self.id),
+            'email': self.email,
+            'last_seen': (
+                None if self.last_seen is None
+                else f'{self.last_seen.isoformat()}Z'
+            ),
+            'member_since': f'{self.member_since.isoformat()}Z',
+            'username': self.username,
+            'full_name': self.full_name,
+            'about_me': self.about_me,
+            'website': self.website,
+            'location': self.location,
+            'organization': self.organization,
+            'job_status_mail_notification_level': \
+                    self.setting_job_status_mail_notification_level.name,
+            'profile_privacy_settings': {
+                'is_public': self.is_public,
+                'show_email': self.has_profile_privacy_setting(ProfilePrivacySettings.SHOW_EMAIL),
+                'show_last_seen': self.has_profile_privacy_setting(ProfilePrivacySettings.SHOW_LAST_SEEN),
+                'show_member_since': self.has_profile_privacy_setting(ProfilePrivacySettings.SHOW_MEMBER_SINCE)
+            }
+        }
+        if backrefs:
+            json_serializeable['role'] = \
+                self.role.to_json_serializeable(backrefs=True)
+        if relationships:
+            json_serializeable['corpus_follower_associations'] = {
+                x.hashid: x.to_json_serializeable()
+                for x in self.corpus_follower_associations
+            }
+            json_serializeable['corpora'] = {
+                x.hashid: x.to_json_serializeable(relationships=True)
+                for x in self.corpora
+            }
+            json_serializeable['jobs'] = {
+                x.hashid: x.to_json_serializeable(relationships=True)
+                for x in self.jobs
+            }
+            json_serializeable['tesseract_ocr_pipeline_models'] = {
+                x.hashid: x.to_json_serializeable(relationships=True)
+                for x in self.tesseract_ocr_pipeline_models
+            }
+            json_serializeable['spacy_nlp_pipeline_models'] = {
+                x.hashid: x.to_json_serializeable(relationships=True)
+                for x in self.spacy_nlp_pipeline_models
+            }
+
+        if filter_by_privacy_settings:
+            if not self.has_profile_privacy_setting(ProfilePrivacySettings.SHOW_EMAIL):
+                json_serializeable.pop('email')
+            if not self.has_profile_privacy_setting(ProfilePrivacySettings.SHOW_LAST_SEEN):
+                json_serializeable.pop('last_seen')
+            if not self.has_profile_privacy_setting(ProfilePrivacySettings.SHOW_MEMBER_SINCE):
+                json_serializeable.pop('member_since')
+        return json_serializeable

app/services/__init__.py

@@ -1,12 +1,11 @@
 from flask import Blueprint
 from flask_login import login_required
-import os
+from pathlib import Path
 import yaml
 
 
-services_file = \
+services_file = Path(__file__).parent / 'services.yml'
-    os.path.join(os.path.dirname(os.path.abspath(__file__)), 'services.yml')
+with services_file.open('r') as f:
-with open(services_file, 'r') as f:
     SERVICES = yaml.safe_load(f)
 
 bp = Blueprint('services', __name__)

app/services/routes.py

@@ -6,6 +6,7 @@ from app import db, hashids
 from app.models import (
     Job,
     JobInput,
+    JobResult,
     JobStatus,
     TesseractOCRPipelineModel,
     SpaCyNLPPipelineModel
@@ -74,6 +75,8 @@ def tesseract_ocr_pipeline():
     version = request.args.get('version', service_manifest['latest_version'])
     if version not in service_manifest['versions']:
         abort(404)
+    job_results = JobResult.query.all()
+    choosable_job_ids = [job_result.job.hashid for job_result in job_results if job_result.job.service == "file-setup-pipeline" and job_result.filename.endswith('.pdf')]
     form = CreateTesseractOCRPipelineJobForm(prefix='create-job-form', version=version)
     if form.is_submitted():
         if not form.validate():
@@ -111,6 +114,7 @@ def tesseract_ocr_pipeline():
     return render_template(
         'services/tesseract_ocr_pipeline.html.j2',
         title=service_manifest['name'],
+        choosable_job_ids=choosable_job_ids,
         form=form,
         tesseract_ocr_pipeline_models=tesseract_ocr_pipeline_models,
         user_tesseract_ocr_pipeline_models_count=user_tesseract_ocr_pipeline_models_count

app/services/services.yml

@@ -59,3 +59,8 @@ spacy-nlp-pipeline:
         - 'encoding_detection'
       publishing_year: 2022
       url: 'https://gitlab.ub.uni-bielefeld.de/sfb1288inf/spacy-nlp-pipeline/-/releases/v0.1.1'
+    0.1.2:
+      methods:
+        - 'encoding_detection'
+      publishing_year: 2024
+      url: 'https://gitlab.ub.uni-bielefeld.de/sfb1288inf/spacy-nlp-pipeline/-/releases/v0.1.2'

app/static/js/resource-lists/job-output-list.js

@@ -0,0 +1,137 @@
+nopaque.resource_lists.JobOutputList = class JobOutputList extends nopaque.resource_lists.ResourceList {
+  static htmlClass = 'job-output-list';
+
+  constructor(listContainerElement, options = {}) {
+    super(listContainerElement, options);
+    this.listjs.list.addEventListener('click', (event) => {this.onClick(event)});
+    this.isInitialized = false;
+    this.userId = listContainerElement.dataset.userId;
+    this.jobOutput = listContainerElement.dataset.jobOutput;
+    this.jobIds = listContainerElement.dataset.jobIds;
+    if (this.userId === undefined) {return;}
+    app.subscribeUser(this.userId).then((response) => {
+      app.socket.on('PATCH', (patch) => {
+        if (this.isInitialized) {this.onPatch(patch);}
+      });
+    });
+    app.getUser(this.userId).then((user) => {
+      let jobIds = JSON.parse(this.jobIds.replace(/'/g, '"'));
+      let job_results = {};
+      for (let jobId of jobIds) {
+        for (let jobResult of Object.values(user.jobs[jobId].results)) {
+          if (jobResult.mimetype === 'application/pdf') {
+            job_results[jobResult.id] = jobResult;
+            job_results[jobResult.id].description = user.jobs[jobId].description;
+            job_results[jobResult.id].title = user.jobs[jobId].title;
+          }
+        }
+      }
+      this.add(Object.values(job_results));
+      this.isInitialized = true;
+    });
+  }
+
+  get item() {
+    return `
+      <tr class="list-item clickable hoverable">
+        <td><span class="title"></span></td>
+        <td><span class="description"></span></td>
+        <td><span class="filename"></span></td>
+        <td class="right-align">
+          <a class="list-action-trigger btn-flat waves-effect waves-light" data-list-action="add"><i class="material-icons">add</i></a>
+        </td>
+      </tr>
+    `.trim();
+  }
+
+  get valueNames() {
+    return [
+      {data: ['id']},
+      {data: ['creation-date']},
+      'title',
+      'description',
+      'filename'
+    ];
+  }
+
+  initListContainerElement() {
+    if (!this.listContainerElement.hasAttribute('id')) {
+      this.listContainerElement.id = nopaque.Utils.generateElementId('job-output-list-');
+    }
+    let listSearchElementId = nopaque.Utils.generateElementId(`${this.listContainerElement.id}-search-`);
+    this.listContainerElement.innerHTML = `
+      <div class="input-field">
+        <i class="material-icons prefix">search</i>
+        <input id="${listSearchElementId}" class="search" type="text"></input>
+        <label for="${listSearchElementId}">Search job output</label>
+      </div>
+      <table>
+        <thead>
+          <tr>
+            <th>Title</th>
+            <th>Description</th>
+            <th>Filename</th>
+            <th></th>
+          </tr>
+        </thead>
+        <tbody class="list"></tbody>
+      </table>
+      <ul class="pagination"></ul>
+    `;
+  }
+
+  mapResourceToValue(jobOutput) {
+    console.log(jobOutput);
+    return {
+      'id': jobOutput.id,
+      'creation-date': jobOutput.creationDate,
+      'title': jobOutput.title,
+      'description': jobOutput.description,
+      'filename': jobOutput.filename
+    };
+  }
+
+  sort() {
+    this.listjs.sort('title', {order: 'asc'});
+  }
+
+  onClick(event) {
+    let listItemElement = event.target.closest('.list-item[data-id]');
+    if (listItemElement === null) {return;}
+    let itemId = listItemElement.dataset.id;
+    let listActionElement = event.target.closest('.list-action-trigger[data-list-action]');
+    let listAction = listActionElement === null ? 'add' : listActionElement.dataset.listAction;
+    switch (listAction) {
+      case 'add': {
+        listActionElement.querySelector('i').textContent = 'done';
+        listActionElement.dataset.listAction = 'remove';
+        break;
+      }
+      case 'remove': {
+        listActionElement.querySelector('i').textContent = 'add';
+        listActionElement.dataset.listAction = 'add';
+        break;
+      }
+      default: {
+        break;
+      }
+    }
+  }
+
+  // onPatch(patch) {
+  //   let re = new RegExp(`^/users/${this.userId}/jobs/${this.jobId}/results/([A-Za-z0-9]*)`);
+  //   let filteredPatch = patch.filter(operation => re.test(operation.path));
+  //   for (let operation of filteredPatch) {
+  //     switch(operation.op) {
+  //       case 'add': {
+  //         let re = new RegExp(`^/users/${this.userId}/jobs/${this.jobId}/results/([A-Za-z0-9]*)$`);
+  //         if (re.test(operation.path)) {this.add(operation.value);}
+  //         break;
+  //       }
+  //       default: {
+  //         break;
+  //       }
+  //     }
+  //   }
+  // }
+};

app/templates/_base/scripts.html.j2

@@ -52,6 +52,7 @@
   'js/resource-lists/job-input-list.js',
   'js/resource-lists/job-list.js',
   'js/resource-lists/job-result-list.js',
+  'js/resource-lists/job-output-list.js',
   'js/resource-lists/public-corpus-list.js',
   'js/resource-lists/public-user-list.js',
   'js/resource-lists/spacy-nlp-pipeline-model-list.js',

app/templates/contributions/spacy_nlp_pipeline_models/create.html.j2

@@ -42,7 +42,7 @@
             {{ form.hidden_tag() }}
             <div class="row">
               <div class="col s12 l5">
-                {{ wtf.render_field(form.spacy_model_file, accept='.tar.gz', placeholder='Choose a .tar.gz file') }}
+                {{ wtf.render_field(form.spacy_model_file, accept='.tar.gz,.whl', placeholder='Choose a .tar.gz or .whl file') }}
               </div>
               <div class="col s12 l7">
                 {{ wtf.render_field(form.title, material_icon='title') }}

app/templates/services/tesseract_ocr_pipeline.html.j2

@@ -37,6 +37,15 @@
 
     <div class="col s12">
       <h2>Submit a job</h2>
+      <div class="card">
+        <div class="card-content">
+          <p>Add an existing file from your workflow or add a new one below.</p>
+          <div class="job-output-list" data-user-id="{{ current_user.hashid}}" data-job-ids="{{ choosable_job_ids }}"></div>
+        </div>
+        <div class="card-action right-align">
+          <a class="waves-effect waves-light btn"><i class="material-icons right">send</i>Submit</a>
+        </div>
+      </div>
       <div class="card">
         <form class="create-job-form" enctype="multipart/form-data" method="POST">
           <div class="card-content">
@@ -51,6 +60,8 @@
               <div class="col s12 l5">
                 {{ wtf.render_field(form.pdf, accept='application/pdf', placeholder='Choose a PDF file') }}
               </div>
+            </div>
+            <div class="row">
               <div class="col s12 l4">
                 <div class="input-field">
                   <i class="material-icons prefix">language</i>

app/users/cli.py

@@ -1,6 +1,4 @@
 from app.models import User
-import os
-import shutil
 from app import db
 from . import bp
 

app/users/routes.py

@@ -7,7 +7,6 @@ from flask import (
 )
 from flask_breadcrumbs import register_breadcrumb
 from flask_login import current_user
-import os
 from app.models import User
 from . import bp
 from .utils import user_dynamic_list_constructor as user_dlc
@@ -40,8 +39,8 @@ def user_avatar(user_id):
     if user.avatar is None:
         return redirect(url_for('static', filename='images/user_avatar.png'))
     return send_from_directory(
-        os.path.dirname(user.avatar.path),
+        user.avatar.path.parent,
-        os.path.basename(user.avatar.path),
+        user.avatar.path.name,
         as_attachment=True,
         attachment_filename=user.avatar.filename,
         mimetype=user.avatar.mimetype

config.py

@@ -1,6 +1,7 @@
 from dotenv import load_dotenv
 from flask import Flask
 from logging.handlers import RotatingFileHandler
+from pathlib import Path
 from werkzeug.middleware.proxy_fix import ProxyFix
 import logging
 import os
@@ -57,8 +58,7 @@ class Config:
 
     ''' # nopaque # '''
     NOPAQUE_ADMIN = os.environ.get('NOPAQUE_ADMIN')
-    NOPAQUE_DATA_DIR = \
+    NOPAQUE_DATA_DIR = Path(os.environ.get('NOPAQUE_DATA_PATH', '/mnt/nopaque'))
-        os.path.abspath(os.environ.get('NOPAQUE_DATA_PATH', '/mnt/nopaque'))
     NOPAQUE_IS_PRIMARY_INSTANCE = \
         os.environ.get('NOPAQUE_IS_PRIMARY_INSTANCE', 'true').lower() == 'true'
     NOPAQUE_MAIL_SUBJECT_PREFIX = '[nopaque]'
@@ -115,7 +115,7 @@ class Config:
     NOPAQUE_READCOOP_USERNAME = os.environ.get('NOPAQUE_READCOOP_USERNAME')
     NOPAQUE_READCOOP_PASSWORD = os.environ.get('NOPAQUE_READCOOP_PASSWORD')
 
-    NOPAQUE_VERSION='1.0.1'
+    NOPAQUE_VERSION='1.0.2'
 
     @staticmethod
     def init_app(app: Flask):