parallex 0.3.1__py3-none-any.whl → 0.3.2__py3-none-any.whl

Sign up to get free protection for your applications and to get access to all the features.
@@ -18,6 +18,7 @@ class OpenAIClient:
18
18
  azure_endpoint=os.getenv("AZURE_API_BASE"),
19
19
  api_key=os.getenv("AZURE_API_KEY"),
20
20
  api_version=os.getenv("AZURE_API_VERSION"),
21
+ timeout=3000,
21
22
  )
22
23
 
23
24
  async def upload(self, file_path: str) -> FileObject:
parallex/parallex.py CHANGED
@@ -106,8 +106,11 @@ async def _prompts_execute(
106
106
  process_semaphore = asyncio.Semaphore(concurrency)
107
107
  prompt_tasks = []
108
108
  for batch in batch_jobs:
109
+ logger.info(
110
+ f"waiting for batch to complete - {batch.id} - {batch.trace_id}"
111
+ )
109
112
  prompt_task = asyncio.create_task(
110
- _wait_and_create_prompt_responses(batch=batch, client=open_ai_client, semaphore=process_semaphore)
113
+ await _wait_and_create_prompt_responses(batch=batch, client=open_ai_client, semaphore=process_semaphore)
111
114
  )
112
115
  prompt_tasks.append(prompt_task)
113
116
  prompt_response_groups = await asyncio.gather(*prompt_tasks)
@@ -174,7 +177,7 @@ async def _execute(
174
177
 
175
178
  pages = [page for batch_pages in page_groups for page in batch_pages]
176
179
  logger.info(f"pages done. total pages- {len(pages)} - {trace_id}")
177
- sorted_pages = sorted(pages, key=lambda x: x.prompt_index)
180
+ sorted_pages = sorted(pages, key=lambda x: x.page_number)
178
181
 
179
182
  # TODO add combined version of MD to output / save to file system
180
183
  callable_output = ParallexCallableOutput(
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.1
2
2
  Name: parallex
3
- Version: 0.3.1
3
+ Version: 0.3.2
4
4
  Summary: PDF to markdown using Azure OpenAI batch processing
5
5
  Home-page: https://github.com/Summed-AI/parallex
6
6
  Author: Jeff Hostetler
@@ -1,6 +1,6 @@
1
1
  parallex/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
2
2
  parallex/ai/batch_processor.py,sha256=O5q_jaIU0VI93p7Riq4aZ_qUiN9Omxp5GOfn0IqEYgo,1361
3
- parallex/ai/open_ai_client.py,sha256=TRH78oYod_EWpp3hjEh097OT7hwsQmtv44_j3X9Frxo,2047
3
+ parallex/ai/open_ai_client.py,sha256=CT8q5XEcMr24RlAOaj3-OezruZLczwPtmUafPQWDYXc,2073
4
4
  parallex/ai/output_processor.py,sha256=Rwp8dkLo4xsqooeBh3Xv-uGVbJMG1JQkwyxdUoOs2tQ,1800
5
5
  parallex/ai/uploader.py,sha256=72SZ3KYQSQL1GI0eCUvRDgDxWhz5vLTU_dRhN7cZg84,5443
6
6
  parallex/file_management/converter.py,sha256=Rj-93LXNl2gCY-XUOCZv7DdCNI2-GyRpS5FobnTqwzo,1111
@@ -15,10 +15,10 @@ parallex/models/parallex_prompts_callable_output.py,sha256=IlNX9627_E8aXWQ-vDBuv
15
15
  parallex/models/prompt_response.py,sha256=LcctuyqwiTHWrZHSahwauMaSBsin5Ws6fQRAzGXTsAA,230
16
16
  parallex/models/raw_file.py,sha256=Nlv6u_jlDCXDgU2_Ff7DRbDCx27pB1NZugNhEoaBMQU,483
17
17
  parallex/models/upload_batch.py,sha256=jrnds9ryXg9drL4TF8TGimMVTCDfKaWsBzFv_ed0i88,2068
18
- parallex/parallex.py,sha256=30DKfec3uHCfyoLj0-wMaSltHn3kLsnW0kJsfFFgktk,8745
18
+ parallex/parallex.py,sha256=SqZBu1QFtSfomJFezteURlsMspvozb3ph4YwjDYXn88,8870
19
19
  parallex/utils/constants.py,sha256=508ieZLZ5kse0T4_QyNJp57Aq0DMNFjjyFlsKa0xtek,366
20
20
  parallex/utils/logger.py,sha256=i3ZZ7YTUmhUStbvVME67F9ffnkLOv5ijm7wVUyJT8Ys,440
21
- parallex-0.3.1.dist-info/LICENSE,sha256=wPwCqGrisXnEcpaUxSO79C2mdOUTbtjhLjyy8mVW6p8,1046
22
- parallex-0.3.1.dist-info/METADATA,sha256=PBeHVqfrMvtGgny4FlkkyYt6nY-Ngbk8Z-6rEhukX_Q,4461
23
- parallex-0.3.1.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
24
- parallex-0.3.1.dist-info/RECORD,,
21
+ parallex-0.3.2.dist-info/LICENSE,sha256=wPwCqGrisXnEcpaUxSO79C2mdOUTbtjhLjyy8mVW6p8,1046
22
+ parallex-0.3.2.dist-info/METADATA,sha256=6LWYbikRUS1jpMPfDXZau1L8y0oNN987KC0CnwBADYs,4461
23
+ parallex-0.3.2.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
24
+ parallex-0.3.2.dist-info/RECORD,,