dddennnisss/trueconf_bot
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix: use yolox model (available on 103) + hi_res + GPU for maximum quality

Browse Source
This commit is contained in:
dddennnisss
2026-07-16 16:15:05 +07:00
parent f8085d9ec2
commit aa0923f4e1
1 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
parse_pol177.py
+3 -3
View File
@@ -1,7 +1,7 @@
#!/usr/bin/env python3 #!/usr/bin/env python3
""" """
Parse PDF document using unstructured-api on 192.168.1.103. Parse PDF document using unstructured-api on 192.168.1.103.
Uses maximum quality settings (hi_res + GPU). Uses maximum quality settings (hi_res + yolox model).
Saves parsed JSON to /tmp/pol177_parsed.json. Saves parsed JSON to /tmp/pol177_parsed.json.
""" """
@@ -13,12 +13,12 @@ PDF_PATH = "/opt/documents_xk/АОХКСибцем/07 Служба Вице-пр
API_URL = "http://192.168.1.103:8005/general/v0/general" API_URL = "http://192.168.1.103:8005/general/v0/general"
def parse_pdf(): def parse_pdf():
print("Parsing PDF with maximum quality (hi_res + GPU)...") print("Parsing PDF with maximum quality (hi_res + yolox)...")
# Maximum quality parameters # Maximum quality parameters
form_data = { form_data = {
'strategy': 'hi_res', 'strategy': 'hi_res',
'hi_res_model_name': 'chipper', 'hi_res_model_name': 'yolox',
'pdf_infer_table_structure': 'true', 'pdf_infer_table_structure': 'true',
'extract_images': 'true', 'extract_images': 'true',
'extract_image_block_types': '["image", "table"]', 'extract_image_block_types': '["image", "table"]',