Separate cluster and matches enpoints (benetech#183)

server/server/api/__init__.py

@@ -1,6 +1,6 @@
 # Disable flake8 issue F401 as we need these imports to configure api
 # but not going to re-export them from the __init__
-from . import scenes, matches, files, errors, videos  # noqa: F401
+from . import scenes, matches, files, errors, videos, cluster  # noqa: F401
 from .blueprint import api
 
 # Explicitly reexport api

server/server/api/cluster.py

@@ -0,0 +1,44 @@
+from http import HTTPStatus
+
+from flask import jsonify, request, abort
+
+from db.access.matches import FileMatchesRequest, MatchesDAO
+from db.schema import Files
+from .blueprint import api
+from .helpers import parse_positive_int, Fields, parse_positive_float, parse_fields
+from ..model import Transform, database
+
+# Optional file fields
+FILE_FIELDS = Fields(Files.exif, Files.signature, Files.meta, Files.scenes)
+
+
+def parse_params(file):
+    """Parse request parameters."""
+    req = FileMatchesRequest(file=file)
+    req.limit = parse_positive_int(request.args, 'limit', 20)
+    req.offset = parse_positive_int(request.args, 'offset', 0)
+    req.hops = parse_positive_int(request.args, 'hops', 1)
+    req.min_distance = parse_positive_float(request.args, 'min_distance', 0.0)
+    req.max_distance = parse_positive_float(request.args, 'max_distance', 1.0)
+    req.preload = parse_fields(request.args, 'include', FILE_FIELDS)
+    return req
+
+
+@api.route('/files/<int:file_id>/cluster', methods=['GET'])
+def list_file_matches(file_id):
+    file = database.session.query(Files).get(file_id)
+
+    # Handle file not found
+    if file is None:
+        abort(HTTPStatus.NOT_FOUND.value, f"File id not found: {file_id}")
+
+    req = parse_params(file)
+    resp = MatchesDAO.list_file_matches(req, database.session)
+
+    include_flags = {field.key: True for field in req.preload}
+    return jsonify({
+        'files': [Transform.file_dict(file, **include_flags) for file in resp.files],
+        'matches': [Transform.match_dict(match) for match in resp.matches],
+        'total': resp.total,
+        'hops': req.hops,
+    })

server/server/api/matches.py

@@ -1,44 +1,46 @@
 from http import HTTPStatus
 
 from flask import jsonify, request, abort
+from sqlalchemy.orm import joinedload
 
-from db.access.matches import FileMatchesRequest, MatchesDAO
-from db.schema import Files
+from db.access.files import FilesDAO
+from db.schema import Matches, Files
 from .blueprint import api
-from .helpers import parse_positive_int, Fields, parse_positive_float, parse_fields
+from .helpers import parse_positive_int, Fields, parse_fields
 from ..model import Transform, database
 
 # Optional file fields
 FILE_FIELDS = Fields(Files.exif, Files.signature, Files.meta, Files.scenes)
 
 
-def parse_params(file):
-    """Parse request parameters."""
-    req = FileMatchesRequest(file=file)
-    req.limit = parse_positive_int(request.args, 'limit', 20)
-    req.offset = parse_positive_int(request.args, 'offset', 0)
-    req.hops = parse_positive_int(request.args, 'hops', 1)
-    req.min_distance = parse_positive_float(request.args, 'min_distance', 0.0)
-    req.max_distance = parse_positive_float(request.args, 'max_distance', 1.0)
-    req.preload = parse_fields(request.args, 'include', FILE_FIELDS)
-    return req
-
-
 @api.route('/files/<int:file_id>/matches', methods=['GET'])
 def list_file_matches(file_id):
+    limit = parse_positive_int(request.args, 'limit', 20)
+    offset = parse_positive_int(request.args, 'offset', 0)
+    include_fields = parse_fields(request.args, 'include', FILE_FIELDS)
+
     file = database.session.query(Files).get(file_id)
 
     # Handle file not found
     if file is None:
         abort(HTTPStatus.NOT_FOUND.value, f"File id not found: {file_id}")
 
-    req = parse_params(file)
-    resp = MatchesDAO.list_file_matches(req, database.session)
+    query = FilesDAO.file_matches(file_id, database.session).options(
+        joinedload(Matches.match_video_file),
+        joinedload(Matches.query_video_file)
+    )
+
+    # Preload file fields
+    query = FILE_FIELDS.preload(query, include_fields, Matches.match_video_file)
+    query = FILE_FIELDS.preload(query, include_fields, Matches.query_video_file)
+
+    # Get requested slice
+    total = query.count()
+    items = query.offset(offset).limit(limit).all()
 
-    include_flags = {field.key: True for field in req.preload}
+    include_flags = {field: True for field in include_fields}
     return jsonify({
-        'files': [Transform.file_dict(file, **include_flags) for file in resp.files],
-        'matches': [Transform.match_dict(match) for match in resp.matches],
-        'total': resp.total,
-        'hops': req.hops,
+        'items': [Transform.file_match_dict(item, file_id, **include_flags) for item in items],
+        'total': total,
+        'offset': offset,
     })

server/server/model.py

@@ -109,6 +109,7 @@ def file_match_dict(match, file_id, *, meta=False, signature=False, scenes=False
         else:
             matched = match.match_video_file
         return {
+            "id": match.id,
             "distance": match.distance,
             "file": Transform.file_dict(matched, meta=meta, signature=signature, scenes=scenes, exif=exif)
         }