apunkt/PageIndex
1
0
Fork 0
mirror of https://github.com/VectifyAI/PageIndex.git synced 2026-05-25 18:55:17 +02:00
Code Issues Projects Releases Packages Wiki Activity

fix config and logger

Browse source
This commit is contained in:
Ray 2025-04-06 14:49:12 +08:00
parent 9daa4101d1
commit e2cf8bb271
3 changed files with 53 additions and 18 deletions
Download patch file Download diff file Expand all files Collapse all files

11
page_index.py
View file

@ -2,13 +2,10 @@ import os
import json
import copy
import math
import sys
import random
sys.path.append('../..')
import re
from utils import *
from .utils import *
import os
from types import SimpleNamespace as config
from concurrent.futures import ThreadPoolExecutor, as_completed
import argparse
@ -1015,6 +1012,8 @@ def tree_parser(page_list, opt, logger=None):
def page_index_main(doc, opt=None):
opt = merge_config(opt, get_default_opt())
logger = JsonLogger(doc)
is_valid_pdf = (
@ -1039,12 +1038,12 @@ def page_index_main(doc, opt=None):
if opt.if_add_doc_description == 'yes':
doc_description = generate_doc_description(structure, model=opt.model)
return {
'doc_name': os.path.basename(doc),
'doc_name': get_pdf_name(doc),
'doc_description': doc_description,
'structure': structure,
}
return {
'doc_name': os.path.basename(doc),
'doc_name': get_pdf_name(doc),
'structure': structure,
}