plastron-web 4.3.2__py3-none-any.whl

plastron/web/__init__.py

@@ -0,0 +1,87 @@
+import logging
+import os
+import yaml
+import urllib.parse
+from pathlib import Path
+from argparse import Namespace
+from flask import Flask, url_for
+from werkzeug.exceptions import NotFound
+
+from plastron.context import PlastronContext
+from plastron.jobs.importjob import ImportJob
+from plastron.jobs import JobError, JobConfigError, JobNotFoundError, Jobs
+from plastron.web.activitystream import activitystream_bp
+from plastron.utils import envsubst
+
+logger = logging.getLogger(__name__)
+
+
+def job_url(job_id):
+    return url_for('show_job', _external=True, job_id=job_id)
+
+
+def items(log):
+    return {
+        'count': len(log),
+        'items': [c for c in log]
+    }
+
+
+def latest_dropped_items(job: ImportJob):
+    latest_run = job.latest_run()
+    if latest_run is None:
+        return {}
+
+    return {
+        'timestamp': latest_run.timestamp,
+        'failed': items(latest_run.failed_items),
+        'invalid': items(latest_run.invalid_items)
+    }
+
+
+def create_app(config_file: str):
+    app = Flask(__name__)
+    with open(config_file, "r") as stream:
+        config = envsubst(yaml.safe_load(stream))
+        app.config['CONTEXT'] = PlastronContext(config=config, args=Namespace(delegated_user=None))
+    jobs_dir = Path(os.environ.get('JOBS_DIR', 'jobs'))
+    jobs = Jobs(directory=jobs_dir)
+    app.register_blueprint(activitystream_bp)
+
+    def get_job(job_id: str):
+        return jobs.get_job(ImportJob, urllib.parse.unquote(job_id))
+
+    @app.route('/jobs')
+    def list_jobs():
+        if not jobs_dir.exists():
+            logger.warning(f'Jobs directory "{jobs_dir.absolute()}" does not exist; returning empty list')
+            return {'jobs': []}
+        job_ids = sorted(f.name for f in jobs_dir.iterdir() if f.is_dir())
+        return {'jobs': [{'@id': job_url(job_id), 'job_id': job_id} for job_id in job_ids]}
+
+    @app.route('/jobs/<path:job_id>')
+    def show_job(job_id):
+        try:
+            job = get_job(job_id)
+            job.load_config()
+        except JobNotFoundError:
+            logger.warning(f'Job {job_id} not found')
+            raise NotFound
+        except JobConfigError:
+            logger.warning(f'Cannot open config file for job {job_id}')
+            # TODO: more complete information in the response body?
+            raise NotFound
+
+        try:
+            return {
+                '@id': job_url(job_id),
+                **vars(job.config),
+                'runs': job.runs,
+                'completed': items(job.completed_log),
+                'dropped': latest_dropped_items(job),
+                'total': job.get_metadata().total
+            }
+        except JobError as e:
+            raise NotFound from e
+
+    return app

plastron/web/activitystream.py

@@ -0,0 +1,94 @@
+import json
+import logging
+from typing import List
+from uuid import uuid4
+
+from flask import Blueprint, current_app, jsonify, request
+from rdflib import Graph
+
+from plastron.namespaces import activitystreams, rdf, umdact
+from plastron.repo.publish import PublishableResource
+
+logger = logging.getLogger(__name__)
+
+activitystream_bp = Blueprint('activitystream', __name__, template_folder='templates')
+
+
+@activitystream_bp.route('/inbox', methods=['POST'])
+def new_activity():
+    try:
+        activity = Activity(from_json=request.get_json())
+        ctx = current_app.config['CONTEXT']
+        for uri in activity.objects:
+            resource: PublishableResource = ctx.repo[uri:PublishableResource].read()
+            if activity.publish:
+                resource.publish(
+                    handle_client=ctx.handle_client,
+                    public_url=ctx.get_public_url(resource),
+                    force_hidden=activity.force_hidden,
+                    force_visible=False,
+                )
+            elif activity.unpublish:
+                resource.unpublish(
+                    force_hidden=activity.force_hidden,
+                    force_visible=False,
+                )
+        return {}, 201
+    except ValidationError as e:
+        logger.error(f'Exception: {e}')
+        return jsonify({'error': str(e)}), 400
+    except Exception as e:
+        logger.error(f'Exception: {e}')
+        return jsonify({'error': str(e)}), 500
+
+
+class Activity:
+    def __init__(self, from_json: str):
+        self.id = uuid4()
+        self._objects = []
+        self._publish = False
+        self._unpublish = False
+        self._force_hidden = False
+        g = Graph()
+        g.parse(data=json.dumps(from_json), format='json-ld')
+
+        for s, p, o in g:
+            if activitystreams.object == p:
+                self._objects.append(str(o))
+            elif rdf.type == p:
+                if o == umdact.Publish:
+                    self._publish = True
+                elif o == umdact.Unpublish:
+                    self._unpublish = True
+                elif o == umdact.PublishHidden:
+                    self._publish = True
+                    self._force_hidden = True
+                else:
+                    raise ValidationError(f'Invalid Activity type: {str(o)}')
+        if not self.publish and not self.unpublish:
+            raise ValidationError(f'Invalid JSON-LD provided: Type not specified.')
+        if not self.objects:
+            raise ValidationError(f'Invalid JSON-LD provided: Object(s) not specified.')
+
+    def __str__(self):
+        return self.id
+
+    @property
+    def publish(self) -> bool:
+        return self._publish
+
+    @property
+    def unpublish(self) -> bool:
+        return self._unpublish
+
+    @property
+    def force_hidden(self) -> bool:
+        return self._force_hidden
+
+    @property
+    def objects(self) -> List[str]:
+        return self._objects
+
+
+class ValidationError(Exception):
+    pass

plastron/web/server.py

@@ -0,0 +1,28 @@
+import click
+from dotenv import load_dotenv
+from waitress import serve
+
+from plastron.web import create_app
+
+
+@click.command()
+@click.option(
+    '--listen',
+    default='0.0.0.0:5000',
+    help='Address and port to listen on. Default is "0.0.0.0:5000".',
+    metavar='[ADDRESS]:PORT',
+)
+@click.option(
+    '-c', '--config-file',
+    type=click.Path(exists=True),
+    help='Configuration file',
+    required=True,
+)
+def run(listen: bool, config_file: str):
+    load_dotenv()
+    app = create_app(config_file)
+    serve(app, listen=listen)
+
+
+if __name__ == "__main__":
+    run()

plastron_web-4.3.2.dist-info/METADATA

@@ -0,0 +1,104 @@
+Metadata-Version: 2.1
+Name: plastron-web
+Version: 4.3.2
+Summary: Plastron HTTP web app
+Author-email: University of Maryland Libraries <lib-ssdr@umd.edu>, Josh Westgard <westgard@umd.edu>, Peter Eichman <peichman@umd.edu>, Mohamed Abdul Rasheed <mohideen@umd.edu>, Ben Wallberg <wallberg@umd.edu>, David Steelman <dsteelma@umd.edu>, Marc Andreu Grillo Aguilar <aguilarm@umd.edu>
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+Requires-Dist: click
+Requires-Dist: flask
+Requires-Dist: plastron-repo
+Requires-Dist: plastron-utils
+Requires-Dist: python-dotenv
+Requires-Dist: waitress
+Provides-Extra: test
+Requires-Dist: pytest ; extra == 'test'
+Requires-Dist: pytest-cov ; extra == 'test'
+
+# plastron-web
+
+HTTP server for synchronous remote operations
+
+## Running with Python
+
+As a Flask application: 
+
+```bash
+flask --app plastron.web:create_app("/path/to/docker-plastron.yml") run
+```
+
+To enable debugging, for hot code reloading, set `FLASK_DEBUG=1` either on 
+the command line or in a `.env` file:
+
+```bash
+FLASK_DEBUG=1 flask --app plastron.web:create_app("/path/to/docker-plastron.yml") run
+```
+
+Using the console script entrypoint, which runs the application with the 
+[Waitress] WSGI server:
+
+```bash
+plastrond-http
+```
+
+## Docker Image
+
+The plastron-stomp package contains a [Dockerfile](Dockerfile) for
+building the `plastrond-http` Docker image.
+
+### Building
+
+**Important:** This image **MUST** be built from the main _plastron_
+project directory, in order to include the other plastron packages in the
+build context.
+
+```bash
+docker build -t docker.lib.umd.edu/plastrond-http:latest \
+    -f plastron-web/Dockerfile .
+```
+
+### Running with Docker Swarm
+
+This repository contains a [compose.yml](compose.yml) file that defines 
+part of a `plastrond` Docker stack intended to be run alongside the
+[umd-fcrepo-docker] stack. This repository's configuration adds a
+`plastrond-http` container.
+
+```bash
+# if you are not already running in swarm mode
+docker swarm init
+
+# build the image
+docker build -t docker.lib.umd.edu/plastrond-http:latest \
+    -f plastron-web/Dockerfile .
+
+# Copy the docker-plastron-template.yml and edit the configuration
+cp docker-plastron.template.yml docker-plastron.yml
+vim docker-plastron.yml
+
+# deploy the stack to run the HTTP webapp
+docker stack deploy -c plastron-web/compose.yml plastrond
+```
+
+To watch the logs:
+
+```bash
+docker service logs -f plastrond_http
+```
+
+To stop the HTTP service:
+
+```bash
+docker service rm plastrond_http
+```
+
+## Configuration
+
+The application is configured through environment variables.
+
+| Name       | Value                                      | Default |
+|------------|--------------------------------------------|---------|
+| `JOBS_DIR` | Root directory for storing job information | `jobs`  |
+
+[umd-fcrepo-docker]: https://github.com/umd-lib/umd-fcrepo-docker
+[Waitress]: https://pypi.org/project/waitress/

plastron_web-4.3.2.dist-info/RECORD

@@ -0,0 +1,8 @@
+plastron/web/__init__.py,sha256=yDyDciDaSibjb2VlloOrZ51zQZ-_uqCzmue91l-7E70,2687
+plastron/web/activitystream.py,sha256=Fsi563cKd3XWGTZ8TZ-W1jyyuTUSinlZJva-bMZJrd4,2956
+plastron/web/server.py,sha256=UXSNsc35pHnvHs4BvFW5-2Y4EFgDlOppPhahkP66A5c,578
+plastron_web-4.3.2.dist-info/METADATA,sha256=gUbvvW9f3IXiIoW5ZiJUp95qLTqZBtKBg6ZgkSheHDs,2896
+plastron_web-4.3.2.dist-info/WHEEL,sha256=R0nc6qTxuoLk7ShA2_Y-UWkN8ZdfDBG2B6Eqpz2WXbs,91
+plastron_web-4.3.2.dist-info/entry_points.txt,sha256=joOgIG9x7DNrL9ngBrJEI6aa8EtqCYS5Nv5wHLvcCFw,59
+plastron_web-4.3.2.dist-info/top_level.txt,sha256=N9Rg78jS1475MaY-sd8HcULh6H1BwDebrZ50iTzBLmI,9
+plastron_web-4.3.2.dist-info/RECORD,,

plastron_web-4.3.2.dist-info/WHEEL

@@ -0,0 +1,5 @@
+Wheel-Version: 1.0
+Generator: setuptools (72.1.0)
+Root-Is-Purelib: true
+Tag: py3-none-any
+

plastron_web-4.3.2.dist-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+plastrond-http = plastron.web.server:run

plastron_web-4.3.2.dist-info/top_level.txt

@@ -0,0 +1 @@
+plastron