parallex 0.3.0__tar.gz → 0.3.2__tar.gz

{parallex-0.3.0 → parallex-0.3.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: parallex
-Version: 0.3.0
+Version: 0.3.2
 Summary: PDF to markdown using Azure OpenAI batch processing
 Home-page: https://github.com/Summed-AI/parallex
 Author: Jeff Hostetler

{parallex-0.3.0 → parallex-0.3.2}/parallex/ai/open_ai_client.py

@@ -18,6 +18,7 @@ class OpenAIClient:
             azure_endpoint=os.getenv("AZURE_API_BASE"),
             api_key=os.getenv("AZURE_API_KEY"),
             api_version=os.getenv("AZURE_API_VERSION"),
+            timeout=3000,
         )
 
     async def upload(self, file_path: str) -> FileObject:

{parallex-0.3.0 → parallex-0.3.2}/parallex/ai/uploader.py

@@ -76,8 +76,8 @@ async def upload_prompts_for_processing(
         jsonl = _simple_jsonl_format(prompt_custom_id, prompt)
         with open(upload_file_location, "a") as jsonl_file:
             jsonl_file.write(json.dumps(jsonl) + "\n")
-        batch_file = await _create_batch_file(client, trace_id, upload_file_location)
-        batch_files.append(batch_file)
+    batch_file = await _create_batch_file(client, trace_id, upload_file_location)
+    batch_files.append(batch_file)
     return batch_files
 
 

{parallex-0.3.0 → parallex-0.3.2}/parallex/parallex.py

@@ -46,6 +46,7 @@ async def parallex(
         )
     except Exception as e:
         logger.error(f"Error occurred: {e}")
+        raise e
     finally:
         await _delete_associated_files(open_ai_client, remote_file_handler)
 
@@ -69,6 +70,7 @@ async def parallex_simple_prompts(
         )
     except Exception as e:
         logger.error(f"Error occurred: {e}")
+        raise e
     finally:
         await _delete_associated_files(open_ai_client, remote_file_handler)
 
@@ -101,30 +103,21 @@ async def _prompts_execute(
             start_batch_tasks.append(batch_task)
         batch_jobs = await asyncio.gather(*start_batch_tasks)
 
+        process_semaphore = asyncio.Semaphore(concurrency)
         prompt_tasks = []
         for batch in batch_jobs:
             logger.info(
                 f"waiting for batch to complete - {batch.id} - {batch.trace_id}"
             )
-            page_task = asyncio.create_task(
-                await wait_for_batch_completion(client=open_ai_client, batch=batch)
-            )
-            prompt_tasks.append(page_task)
-
-        output_file_ids = await asyncio.gather(*prompt_tasks)
-
-        prompts_output = []
-        for output_file_id in output_file_ids:
-            logger.info(f"batch completed - {batch.id} - {batch.trace_id}")
-            prompts_output.append(
-                await process_prompts_output(
-                    client=open_ai_client, output_file_id=output_file_id
-                )
+            prompt_task = asyncio.create_task(
+                await _wait_and_create_prompt_responses(batch=batch, client=open_ai_client, semaphore=process_semaphore)
             )
+            prompt_tasks.append(prompt_task)
+        prompt_response_groups = await asyncio.gather(*prompt_tasks)
 
-        flat_prompts = [page for batch in prompts_output for page in batch]
+        flat_responses = [response for batch in prompt_response_groups for response in batch]
 
-        sorted_responses = sorted(flat_prompts, key=lambda x: x.prompt_index)
+        sorted_responses = sorted(flat_responses, key=lambda x: x.prompt_index)
         callable_output = ParallexPromptsCallableOutput(
             original_prompts=prompts,
             trace_id=trace_id,
@@ -211,6 +204,19 @@ async def _wait_and_create_pages(
         return page_responses
 
 
+async def _wait_and_create_prompt_responses(
+    batch: UploadBatch, client: OpenAIClient, semaphore: asyncio.Semaphore
+):
+    async with semaphore:
+        logger.info(f"waiting for batch to complete - {batch.id} - {batch.trace_id}")
+        output_file_id = await wait_for_batch_completion(client=client, batch=batch)
+        logger.info(f"batch completed - {batch.id} - {batch.trace_id}")
+        prompt_responses = await process_prompts_output(
+            client=client, output_file_id=output_file_id
+        )
+        return prompt_responses
+
+
 async def _create_batch_jobs(
     batch_file: BatchFile,
     client: OpenAIClient,

{parallex-0.3.0 → parallex-0.3.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "parallex"
-version = "0.3.0"
+version = "0.3.2"
 description = "PDF to markdown using Azure OpenAI batch processing"
 authors = ["Jeff Hostetler <jeff@summed.ai>", "Kevin Bao <kevin@summed.ai>"]
 repository = "https://github.com/Summed-AI/parallex"