Filter intentaire results by confidence

bookwyrm/connectors/abstract_connector.py

@@ -48,20 +48,18 @@ class AbstractMinimalConnector(ABC):
         # searched as free text. This, instead, only searches isbn if it's isbn-y
         return f"{self.search_url}{query}"
 
-    def process_search_response(self, query, data):
+    def process_search_response(self, query, data, min_confidence):
         """Format the search results based on the formt of the query"""
-        # TODO: inventaire min confidence
-        parser = self.parse_search_data
         if maybe_isbn(query):
-            parser = self.parse_isbn_search_data
-        return list(parser(data))[:10]
+            return list(self.parse_isbn_search_data(data))[:10]
+        return list(self.parse_search_data(data, min_confidence))[:10]
 
     @abstractmethod
     def get_or_create_book(self, remote_id):
         """pull up a book record by whatever means possible"""
 
     @abstractmethod
-    def parse_search_data(self, data):
+    def parse_search_data(self, data, min_confidence):
         """turn the result json from a search into a list"""
 
     @abstractmethod

bookwyrm/connectors/bookwyrm_connector.py

@@ -10,7 +10,7 @@ class Connector(AbstractMinimalConnector):
     def get_or_create_book(self, remote_id):
         return activitypub.resolve_remote_id(remote_id, model=models.Edition)
 
-    def parse_search_data(self, data):
+    def parse_search_data(self, data, min_confidence):
         for search_result in data:
             search_result["connector"] = self
             yield SearchResult(**search_result)

bookwyrm/connectors/connector_manager.py

@@ -22,7 +22,7 @@ class ConnectorException(HTTPError):
     """when the connector can't do what was asked"""
 
 
-async def get_results(session, url, params, query, connector):
+async def get_results(session, url, min_confidence, query, connector):
     """try this specific connector"""
     # pylint: disable=line-too-long
     headers = {
@@ -31,6 +31,7 @@ async def get_results(session, url, params, query, connector):
         ),
         "User-Agent": USER_AGENT,
     }
+    params = {"min_confidence": min_confidence}
     try:
         async with session.get(url, headers=headers, params=params) as response:
             if not response.ok:
@@ -45,7 +46,7 @@ async def get_results(session, url, params, query, connector):
 
             return {
                 "connector": connector,
-                "results": connector.process_search_response(query, raw_data),
+                "results": connector.process_search_response(query, raw_data, min_confidence),
             }
     except asyncio.TimeoutError:
         logger.info("Connection timed out for url: %s", url)
@@ -53,7 +54,7 @@ async def get_results(session, url, params, query, connector):
         logger.exception(err)
 
 
-async def async_connector_search(query, items, params):
+async def async_connector_search(query, items, min_confidence):
     """Try a number of requests simultaneously"""
     timeout = aiohttp.ClientTimeout(total=SEARCH_TIMEOUT)
     async with aiohttp.ClientSession(timeout=timeout) as session:
@@ -61,7 +62,7 @@ async def async_connector_search(query, items, params):
         for url, connector in items:
             tasks.append(
                 asyncio.ensure_future(
-                    get_results(session, url, params, query, connector)
+                    get_results(session, url, min_confidence, query, connector)
                 )
             )
 
@@ -87,8 +88,7 @@ def search(query, min_confidence=0.1, return_first=False):
         items.append((url, connector))
 
     # load as many results as we can
-    params = {"min_confidence": min_confidence}
-    results = asyncio.run(async_connector_search(query, items, params))
+    results = asyncio.run(async_connector_search(query, items, min_confidence))
 
     if return_first:
         # find the best result from all the responses and return that

bookwyrm/connectors/inventaire.py

@@ -77,13 +77,15 @@ class Connector(AbstractConnector):
             **{k: data.get(k) for k in ["uri", "image", "labels", "sitelinks", "type"]},
         }
 
-    def parse_search_data(self, data):
+    def parse_search_data(self, data, min_confidence):
         for search_result in data.get("results", []):
             images = search_result.get("image")
             cover = f"{self.covers_url}/img/entities/{images[0]}" if images else None
             # a deeply messy translation of inventaire's scores
             confidence = float(search_result.get("_score", 0.1))
             confidence = 0.1 if confidence < 150 else 0.999
+            if confidence < min_confidence:
+                continue
             yield SearchResult(
                 title=search_result.get("label"),
                 key=self.get_remote_id(search_result.get("uri")),

bookwyrm/connectors/openlibrary.py

@@ -152,7 +152,7 @@ class Connector(AbstractConnector):
         image_name = f"{cover_id}-{size}.jpg"
         return f"{self.covers_url}/b/id/{image_name}"
 
-    def parse_search_data(self, data):
+    def parse_search_data(self, data, min_confidence):
         for search_result in data.get("docs"):
             # build the remote id from the openlibrary key
             key = self.books_url + search_result["key"]