Refactor some loops into list comprehensions (#1185)

langchain/cache.py

@@ -75,9 +75,7 @@ class SQLAlchemyCache(BaseCache):
             .order_by(self.cache_schema.idx)
         )
         with Session(self.engine) as session:
-            generations = []
+            generations = [Generation(text=row[0]) for row in session.execute(stmt)]
-            for row in session.execute(stmt):
-                generations.append(Generation(text=row[0]))
             if len(generations) > 0:
                 return generations
         return None

langchain/chains/llm.py

@@ -124,12 +124,11 @@ class LLMChain(Chain, BaseModel):
 
     def create_outputs(self, response: LLMResult) -> List[Dict[str, str]]:
         """Create outputs from response."""
-        outputs = []
+        return [
-        for generation in response.generations:
             # Get the text of the top generated string.
-            response_str = generation[0].text
+            {self.output_key: generation[0].text}
-            outputs.append({self.output_key: response_str})
+            for generation in response.generations
-        return outputs
+        ]
 
     def _call(self, inputs: Dict[str, Any]) -> Dict[str, str]:
         return self.apply([inputs])[0]
@@ -188,11 +187,9 @@ class LLMChain(Chain, BaseModel):
         self, result: List[Dict[str, str]]
     ) -> Sequence[Union[str, List[str], Dict[str, str]]]:
         if self.prompt.output_parser is not None:
-            new_result = []
+            return [
-            for res in result:
+                self.prompt.output_parser.parse(res[self.output_key]) for res in result
-                text = res[self.output_key]
+            ]
-                new_result.append(self.prompt.output_parser.parse(text))
-            return new_result
         else:
             return result
 

langchain/document_loaders/googledrive.py

@@ -116,22 +116,19 @@ class GoogleDriveLoader(BaseLoader, BaseModel):
         )
         items = results.get("files", [])
 
-        docs = []
+        return [
-        for item in items:
+            self._load_document_from_id(item["id"])
+            for item in items
             # Only support Google Docs for now
-            if item["mimeType"] == "application/vnd.google-apps.document":
+            if item["mimeType"] == "application/vnd.google-apps.document"
-                docs.append(self._load_document_from_id(item["id"]))
+        ]
-        return docs
 
     def _load_documents_from_ids(self) -> List[Document]:
         """Load documents from a list of IDs."""
         if not self.document_ids:
             raise ValueError("document_ids must be set")
 
-        docs = []
+        return [self._load_document_from_id(doc_id) for doc_id in self.document_ids]
-        for doc_id in self.document_ids:
-            docs.append(self._load_document_from_id(doc_id))
-        return docs
 
     def load(self) -> List[Document]:
         """Load documents."""

langchain/document_loaders/hn.py

@@ -30,12 +30,13 @@ class HNLoader(WebBaseLoader):
         """Load comments from a HN post."""
         comments = soup_info.select("tr[class='athing comtr']")
         title = soup_info.select_one("tr[id='pagespace']").get("title")
-        documents = []
+        return [
-        for comment in comments:
+            Document(
-            text = comment.text.strip()
+                page_content=comment.text.strip(),
-            metadata = {"source": self.web_path, "title": title}
+                metadata={"source": self.web_path, "title": title},
-            documents.append(Document(page_content=text, metadata=metadata))
+            )
-        return documents
+            for comment in comments
+        ]
 
     def load_results(self, soup: Any) -> List[Document]:
         """Load items from an HN page."""

langchain/document_loaders/paged_pdf.py

@@ -25,12 +25,12 @@ class PagedPDFSplitter(BaseLoader):
         """Load given path as pages."""
         import pypdf
 
-        pdf_file_obj = open(self._file_path, "rb")
+        with open(self._file_path, "rb") as pdf_file_obj:
-        pdf_reader = pypdf.PdfReader(pdf_file_obj)
+            pdf_reader = pypdf.PdfReader(pdf_file_obj)
-        docs = []
+            return [
-        for i, page in enumerate(pdf_reader.pages):
+                Document(
-            text = page.extract_text()
+                    page_content=page.extract_text(),
-            metadata = {"source": self._file_path, "page": i}
+                    metadata={"source": self._file_path, "page": i},
-            docs.append(Document(page_content=text, metadata=metadata))
+                )
-        pdf_file_obj.close()
+                for i, page in enumerate(pdf_reader.pages)
-        return docs
+            ]

langchain/embeddings/huggingface.py

@@ -121,9 +121,7 @@ class HuggingFaceInstructEmbeddings(BaseModel, Embeddings):
         Returns:
             List of embeddings, one for each text.
         """
-        instruction_pairs = []
+        instruction_pairs = [[self.embed_instruction, text] for text in texts]
-        for text in texts:
-            instruction_pairs.append([self.embed_instruction, text])
         embeddings = self.client.encode(instruction_pairs)
         return embeddings.tolist()
 

langchain/evaluation/qa/eval_chain.py

@@ -48,13 +48,13 @@ class QAEvalChain(LLMChain):
         prediction_key: str = "result",
     ) -> List[dict]:
         """Evaluate question answering examples and predictions."""
-        inputs = []
+        inputs = [
-        for i, example in enumerate(examples):
+            {
-            _input = {
                 "query": example[question_key],
                 "answer": example[answer_key],
                 "result": predictions[i][prediction_key],
             }
-            inputs.append(_input)
+            for i, example in enumerate(examples)
+        ]
 
         return self.apply(inputs)

langchain/llms/openai.py

@@ -329,7 +329,7 @@ class BaseOpenAI(BaseLLM, BaseModel):
     ) -> LLMResult:
         """Create the LLMResult from the choices and prompts."""
         generations = []
-        for i, prompt in enumerate(prompts):
+        for i, _ in enumerate(prompts):
             sub_choices = choices[i * self.n : (i + 1) * self.n]
             generations.append(
                 [

langchain/utilities/searx_search.py

@@ -304,7 +304,6 @@ class SearxSearchWrapper(BaseModel):
 
 
         """
-        metadata_results = []
         _params = {
             "q": query,
         }
@@ -314,14 +313,14 @@ class SearxSearchWrapper(BaseModel):
         results = self._searx_api_query(params).results[:num_results]
         if len(results) == 0:
             return [{"Result": "No good Search Result was found"}]
-        for result in results:
+
-            metadata_result = {
+        return [
+            {
                 "snippet": result.get("content", ""),
                 "title": result["title"],
                 "link": result["url"],
                 "engines": result["engines"],
                 "category": result["category"],
             }
-            metadata_results.append(metadata_result)
+            for result in results
-
+        ]
-        return metadata_results