parallex 0.3.0__py3-none-any.whl → 0.3.1__py3-none-any.whl

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/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 +103,18 @@ 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_task = asyncio.create_task(
+                _wait_and_create_prompt_responses(batch=batch, client=open_ai_client, semaphore=process_semaphore)
             )
-            prompt_tasks.append(page_task)
+            prompt_tasks.append(prompt_task)
+        prompt_response_groups = await asyncio.gather(*prompt_tasks)
 
-        output_file_ids = await asyncio.gather(*prompt_tasks)
+        flat_responses = [response for batch in prompt_response_groups for response in batch]
 
-        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
-                )
-            )
-
-        flat_prompts = [page for batch in prompts_output for page 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,
@@ -184,7 +174,7 @@ async def _execute(
 
         pages = [page for batch_pages in page_groups for page in batch_pages]
         logger.info(f"pages done. total pages- {len(pages)} - {trace_id}")
-        sorted_pages = sorted(pages, key=lambda x: x.page_number)
+        sorted_pages = sorted(pages, key=lambda x: x.prompt_index)
 
         # TODO add combined version of MD to output / save to file system
         callable_output = ParallexCallableOutput(
@@ -211,6 +201,19 @@ 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.dist-info → parallex-0.3.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: parallex
-Version: 0.3.0
+Version: 0.3.1
 Summary: PDF to markdown using Azure OpenAI batch processing
 Home-page: https://github.com/Summed-AI/parallex
 Author: Jeff Hostetler

{parallex-0.3.0.dist-info → parallex-0.3.1.dist-info}/RECORD

@@ -2,7 +2,7 @@ parallex/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 parallex/ai/batch_processor.py,sha256=O5q_jaIU0VI93p7Riq4aZ_qUiN9Omxp5GOfn0IqEYgo,1361
 parallex/ai/open_ai_client.py,sha256=TRH78oYod_EWpp3hjEh097OT7hwsQmtv44_j3X9Frxo,2047
 parallex/ai/output_processor.py,sha256=Rwp8dkLo4xsqooeBh3Xv-uGVbJMG1JQkwyxdUoOs2tQ,1800
-parallex/ai/uploader.py,sha256=9GvrzuaQAxqRiYN5dUHWjFeIFXezH0Y7ARnzBkEHbL0,5451
+parallex/ai/uploader.py,sha256=72SZ3KYQSQL1GI0eCUvRDgDxWhz5vLTU_dRhN7cZg84,5443
 parallex/file_management/converter.py,sha256=Rj-93LXNl2gCY-XUOCZv7DdCNI2-GyRpS5FobnTqwzo,1111
 parallex/file_management/file_finder.py,sha256=BPvrkxZlwOYmRXzzS138wGTsVzuhDIKfQZn0CISUj3o,1598
 parallex/file_management/remote_file_handler.py,sha256=jsI9NhOrKQR8K3yo536lGplVBGis9XY0G4dRpumgWFM,213
@@ -15,10 +15,10 @@ parallex/models/parallex_prompts_callable_output.py,sha256=IlNX9627_E8aXWQ-vDBuv
 parallex/models/prompt_response.py,sha256=LcctuyqwiTHWrZHSahwauMaSBsin5Ws6fQRAzGXTsAA,230
 parallex/models/raw_file.py,sha256=Nlv6u_jlDCXDgU2_Ff7DRbDCx27pB1NZugNhEoaBMQU,483
 parallex/models/upload_batch.py,sha256=jrnds9ryXg9drL4TF8TGimMVTCDfKaWsBzFv_ed0i88,2068
-parallex/parallex.py,sha256=7YFKnKOkFHoTC7CCHhrXG1JTxprbvw0QkNGOEPYJbvQ,8500
+parallex/parallex.py,sha256=30DKfec3uHCfyoLj0-wMaSltHn3kLsnW0kJsfFFgktk,8745
 parallex/utils/constants.py,sha256=508ieZLZ5kse0T4_QyNJp57Aq0DMNFjjyFlsKa0xtek,366
 parallex/utils/logger.py,sha256=i3ZZ7YTUmhUStbvVME67F9ffnkLOv5ijm7wVUyJT8Ys,440
-parallex-0.3.0.dist-info/LICENSE,sha256=wPwCqGrisXnEcpaUxSO79C2mdOUTbtjhLjyy8mVW6p8,1046
-parallex-0.3.0.dist-info/METADATA,sha256=hIIhGrV5PE-E-lkWf-kBE3QBPevKSVRHkw0hUx_iqik,4461
-parallex-0.3.0.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
-parallex-0.3.0.dist-info/RECORD,,
+parallex-0.3.1.dist-info/LICENSE,sha256=wPwCqGrisXnEcpaUxSO79C2mdOUTbtjhLjyy8mVW6p8,1046
+parallex-0.3.1.dist-info/METADATA,sha256=PBeHVqfrMvtGgny4FlkkyYt6nY-Ngbk8Z-6rEhukX_Q,4461
+parallex-0.3.1.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
+parallex-0.3.1.dist-info/RECORD,,