spiderforce4ai 2.5.7__py3-none-any.whl → 2.5.8__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- spiderforce4ai/post_extraction_agent.py +12 -0
- {spiderforce4ai-2.5.7.dist-info → spiderforce4ai-2.5.8.dist-info}/METADATA +1 -1
- spiderforce4ai-2.5.8.dist-info/RECORD +7 -0
- spiderforce4ai-2.5.7.dist-info/RECORD +0 -7
- {spiderforce4ai-2.5.7.dist-info → spiderforce4ai-2.5.8.dist-info}/WHEEL +0 -0
- {spiderforce4ai-2.5.7.dist-info → spiderforce4ai-2.5.8.dist-info}/entry_points.txt +0 -0
- {spiderforce4ai-2.5.7.dist-info → spiderforce4ai-2.5.8.dist-info}/top_level.txt +0 -0
@@ -167,6 +167,8 @@ class PostExtractionAgent:
|
|
167
167
|
backup_path = self.config.output_file.with_suffix(f".bak_{int(time.time())}")
|
168
168
|
self.config.output_file.rename(backup_path)
|
169
169
|
self.config.output_file.touch()
|
170
|
+
# Initialize empty results dictionary
|
171
|
+
self.results = {}
|
170
172
|
|
171
173
|
def _process_single_content(self, url: str, content: str) -> Optional[Dict]:
|
172
174
|
"""Process a single piece of content through the LLM."""
|
@@ -259,8 +261,18 @@ class PostExtractionAgent:
|
|
259
261
|
try:
|
260
262
|
if self.config.combine_output and self.config.output_file:
|
261
263
|
self.results[url] = result
|
264
|
+
# Save combined results
|
262
265
|
with open(self.config.output_file, 'w') as f:
|
263
266
|
json.dump(self.results, f, indent=2)
|
267
|
+
|
268
|
+
# Cleanup backup files after successful save
|
269
|
+
for backup_file in self.config.output_file.parent.glob(f"{self.config.output_file.stem}.bak_*"):
|
270
|
+
try:
|
271
|
+
backup_file.unlink()
|
272
|
+
logger.info(f"Cleaned up backup file: {backup_file}")
|
273
|
+
except Exception as e:
|
274
|
+
logger.warning(f"Failed to remove backup file {backup_file}: {e}")
|
275
|
+
|
264
276
|
elif not self.config.combine_output and self.config.output_file:
|
265
277
|
individual_file = self.config.output_file.parent / f"{url.replace('/', '_')}.json"
|
266
278
|
with open(individual_file, 'w') as f:
|
@@ -0,0 +1,7 @@
|
|
1
|
+
spiderforce4ai/__init__.py,sha256=BlrRIrtpDUHjmDedqgXP1KbAAsAH9vwFPncUR5VGGyM,44804
|
2
|
+
spiderforce4ai/post_extraction_agent.py,sha256=nyc6V21E3laTetFG2aEBMmFiXv5goumo8eHoupp3mKA,16430
|
3
|
+
spiderforce4ai-2.5.8.dist-info/METADATA,sha256=eitdlF0yEzoFabYliZWvNSx7hGQMBDVvKwgCh21lU18,9012
|
4
|
+
spiderforce4ai-2.5.8.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
|
5
|
+
spiderforce4ai-2.5.8.dist-info/entry_points.txt,sha256=ibARQxOlDiL1ho12zbDZt4Uq5RKSIk_qk159ZlZ46hc,59
|
6
|
+
spiderforce4ai-2.5.8.dist-info/top_level.txt,sha256=Kth7A21Js7DCp0j5XBBi-FE45SCLouZkeNZU__Yr9Yk,15
|
7
|
+
spiderforce4ai-2.5.8.dist-info/RECORD,,
|
@@ -1,7 +0,0 @@
|
|
1
|
-
spiderforce4ai/__init__.py,sha256=BlrRIrtpDUHjmDedqgXP1KbAAsAH9vwFPncUR5VGGyM,44804
|
2
|
-
spiderforce4ai/post_extraction_agent.py,sha256=e7_f5Fd1BYCvqYAQ9D2jypVt8ScXw55FkX9SOeRA8Co,15812
|
3
|
-
spiderforce4ai-2.5.7.dist-info/METADATA,sha256=F4raYf8aoAlbV7f-YFkCNWfoOlhv2q7voSLzQF9k6s0,9012
|
4
|
-
spiderforce4ai-2.5.7.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
|
5
|
-
spiderforce4ai-2.5.7.dist-info/entry_points.txt,sha256=ibARQxOlDiL1ho12zbDZt4Uq5RKSIk_qk159ZlZ46hc,59
|
6
|
-
spiderforce4ai-2.5.7.dist-info/top_level.txt,sha256=Kth7A21Js7DCp0j5XBBi-FE45SCLouZkeNZU__Yr9Yk,15
|
7
|
-
spiderforce4ai-2.5.7.dist-info/RECORD,,
|
File without changes
|
File without changes
|
File without changes
|