spiderforce4ai 2.5.4__tar.gz → 2.5.6__tar.gz
Sign up to get free protection for your applications and to get access to all the features.
- {spiderforce4ai-2.5.4 → spiderforce4ai-2.5.6}/PKG-INFO +1 -1
- {spiderforce4ai-2.5.4 → spiderforce4ai-2.5.6}/pyproject.toml +1 -1
- {spiderforce4ai-2.5.4 → spiderforce4ai-2.5.6}/setup.py +1 -1
- {spiderforce4ai-2.5.4 → spiderforce4ai-2.5.6}/spiderforce4ai/__init__.py +8 -1
- {spiderforce4ai-2.5.4 → spiderforce4ai-2.5.6}/spiderforce4ai/post_extraction_agent.py +5 -2
- {spiderforce4ai-2.5.4 → spiderforce4ai-2.5.6}/spiderforce4ai.egg-info/PKG-INFO +1 -1
- {spiderforce4ai-2.5.4 → spiderforce4ai-2.5.6}/README.md +0 -0
- {spiderforce4ai-2.5.4 → spiderforce4ai-2.5.6}/setup.cfg +0 -0
- {spiderforce4ai-2.5.4 → spiderforce4ai-2.5.6}/spiderforce4ai.egg-info/SOURCES.txt +0 -0
- {spiderforce4ai-2.5.4 → spiderforce4ai-2.5.6}/spiderforce4ai.egg-info/dependency_links.txt +0 -0
- {spiderforce4ai-2.5.4 → spiderforce4ai-2.5.6}/spiderforce4ai.egg-info/entry_points.txt +0 -0
- {spiderforce4ai-2.5.4 → spiderforce4ai-2.5.6}/spiderforce4ai.egg-info/not-zip-safe +0 -0
- {spiderforce4ai-2.5.4 → spiderforce4ai-2.5.6}/spiderforce4ai.egg-info/requires.txt +0 -0
- {spiderforce4ai-2.5.4 → spiderforce4ai-2.5.6}/spiderforce4ai.egg-info/top_level.txt +0 -0
@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
|
|
4
4
|
|
5
5
|
[project]
|
6
6
|
name = "spiderforce4ai"
|
7
|
-
version = "2.5.
|
7
|
+
version = "2.5.6"
|
8
8
|
description = "Python wrapper for SpiderForce4AI HTML-to-Markdown conversion service with LLM post-processing"
|
9
9
|
readme = "README.md"
|
10
10
|
authors = [
|
@@ -7,7 +7,7 @@ with open("README.md", encoding="utf-8") as f:
|
|
7
7
|
|
8
8
|
setup(
|
9
9
|
name="spiderforce4ai",
|
10
|
-
version="2.5.
|
10
|
+
version="2.5.6",
|
11
11
|
author="Piotr Tamulewicz",
|
12
12
|
author_email="pt@petertam.pro",
|
13
13
|
description="Python wrapper for SpiderForce4AI HTML-to-Markdown conversion service with LLM post-processing",
|
@@ -750,7 +750,10 @@ class SpiderForce4AI:
|
|
750
750
|
for result in successful_results:
|
751
751
|
try:
|
752
752
|
# Process content synchronously since it's not an async method
|
753
|
-
|
753
|
+
extraction_result = agent.process_content(result.url, result.markdown)
|
754
|
+
if extraction_result:
|
755
|
+
result.extraction_result = extraction_result
|
756
|
+
logger.info(f"Successfully processed and transformed content for {result.url}")
|
754
757
|
progress.update(llm_task, advance=1)
|
755
758
|
except Exception as e:
|
756
759
|
console.print(f"[red]Error in post-extraction processing for {result.url}: {str(e)}[/red]")
|
@@ -889,7 +892,9 @@ class SpiderForce4AI:
|
|
889
892
|
# Apply transformation if provided
|
890
893
|
if config.post_agent_transformer_function:
|
891
894
|
try:
|
895
|
+
logger.info(f"Starting transformer function execution for {result.url}")
|
892
896
|
result.extraction_result = config.post_agent_transformer_function(llm_response)
|
897
|
+
logger.info(f"Successfully executed transformer function for {result.url}")
|
893
898
|
except KeyError as e:
|
894
899
|
# Log missing field but continue with transformation
|
895
900
|
missing_field = str(e).strip("'")
|
@@ -897,7 +902,9 @@ class SpiderForce4AI:
|
|
897
902
|
# Add missing field with empty value
|
898
903
|
llm_response[missing_field] = ""
|
899
904
|
# Retry transformation with added field
|
905
|
+
logger.info(f"Retrying transformer function for {result.url} after adding missing field")
|
900
906
|
result.extraction_result = config.post_agent_transformer_function(llm_response)
|
907
|
+
logger.info(f"Successfully executed transformer function on retry for {result.url}")
|
901
908
|
except Exception as e:
|
902
909
|
logger.error(f"Transformer error for {result.url}: {str(e)}")
|
903
910
|
result.extraction_result = llm_response # Use original response if transform fails
|
@@ -306,8 +306,11 @@ class PostExtractionAgent:
|
|
306
306
|
logger.warning(f"Missing field '{field}' in LLM response for {url}, setting to None")
|
307
307
|
result[field] = None
|
308
308
|
|
309
|
-
|
310
|
-
|
309
|
+
logger.info(f"Executing custom transformer function for {url}")
|
310
|
+
transformed_result = self.config.custom_transform_function(result)
|
311
|
+
logger.info(f"Successfully applied custom transformation for {url}")
|
312
|
+
logger.info(f"Webhook response sent for {url}")
|
313
|
+
return transformed_result # Return the transformed result
|
311
314
|
except Exception as e:
|
312
315
|
error_msg = f"Warning: Issue in custom transform for {url}: {str(e)}"
|
313
316
|
logger.warning(error_msg)
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|