deposit: PDF metadata extraction

* Not extracts metadata on upload anymore.
* Provides a REST endpoint for extracting metadata.

Co-Authored-by: Sébastien Délèze <sebastien.deleze@rero.ch>

sonar/modules/deposits/rest.py

@@ -70,16 +70,6 @@ def post(pid=None):
         deposit.files[key]['category'] = request.args['type']
         deposit.files[key]['file_type'] = 'file'
 
-        # Extract data from pdf and populate deposit
-        if request.args['type'] == 'main':
-            pdf_extractor = PDFExtractor()
-            pdf_metadata = format_extracted_data(
-                pdf_extractor.process_raw(request.get_data()))
-
-            # deposit.populate_with_pdf_metadata(
-            #     pdf_metadata, "Deposit #{pid}".format(pid=pid))
-            deposit.files[key]['pdf_metadata'] = pdf_metadata
-
         deposit.commit()
         db.session.commit()
 
@@ -230,3 +220,30 @@ def review(pid=None):
     db.session.commit()
 
     return make_response(jsonify(deposit))
+
+
+@blueprint.route('/extract-pdf-metadata', methods=['GET'])
+def extract_metadata(pid=None):
+    """Publish a deposit or send a message for review."""
+    deposit = DepositRecord.get_record_by_pid(pid)
+
+    if not deposit:
+        abort(400)
+
+    main_file = [
+        file for file in deposit.files if file['category'] == 'main' and
+        file.mimetype == 'application/pdf'
+    ]
+
+    if not main_file:
+        abort(500)
+
+    # Get file content
+    with main_file[0].file.storage().open() as pdf_file:
+        content = pdf_file.read()
+
+    # Extract data from pdf
+    pdf_extractor = PDFExtractor()
+    pdf_metadata = format_extracted_data(pdf_extractor.process_raw(content))
+
+    return make_response(jsonify(pdf_metadata))

tests/api/deposits/test_deposits_rest.py

@@ -134,3 +134,29 @@ def test_review(client, db, db_user_fixture, db_moderator_fixture,
                            }),
                            headers=headers)
     assert response.status_code == 200
+
+
+def test_extract_metadata(app, client, deposit_fixture):
+    """Test PDF metadata extraction."""
+    url = 'https://localhost:5000/deposits/{pid}/extract-pdf-metadata'.format(
+        pid=deposit_fixture['pid'])
+
+    headers = {
+        'Content-Type': 'application/json',
+        'Accept': 'application/json'
+    }
+
+    response = client.get(url, headers=headers)
+    assert response.status_code == 200
+    assert response.json[
+        'title'] == 'High-harmonic generation in quantum spin systems'
+
+    deposit_fixture.files['main.pdf'].remove()
+    response = client.get(url, headers=headers)
+    assert response.status_code == 500
+
+    response = client.get(
+        'https://localhost:5000/deposits/{pid}/extract-pdf-metadata'.format(
+            pid='not-existing'),
+        headers=headers)
+    assert response.status_code == 400

tests/conftest.py

@@ -444,8 +444,9 @@ def deposit_fixture(app, db, db_user_fixture, pdf_file,
     deposit.files['additional.pdf']['category'] = 'additional'
     deposit.files['additional.pdf']['file_type'] = 'file'
 
-    db.session.commit()
+    deposit.commit()
     deposit.reindex()
+    db.session.commit()
 
     current_search.flush_and_refresh('deposits')