indexify 0.0.41__py3-none-any.whl → 0.0.42__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.
- indexify/extractors/pdf_parser.py +1 -3
- {indexify-0.0.41.dist-info → indexify-0.0.42.dist-info}/METADATA +1 -1
- {indexify-0.0.41.dist-info → indexify-0.0.42.dist-info}/RECORD +5 -5
- {indexify-0.0.41.dist-info → indexify-0.0.42.dist-info}/LICENSE.txt +0 -0
- {indexify-0.0.41.dist-info → indexify-0.0.42.dist-info}/WHEEL +0 -0
|
@@ -2,7 +2,6 @@ import tempfile
|
|
|
2
2
|
from enum import Enum
|
|
3
3
|
from typing import List, Optional
|
|
4
4
|
|
|
5
|
-
import deepdoctection as dd
|
|
6
5
|
from pydantic import BaseModel
|
|
7
6
|
|
|
8
7
|
|
|
@@ -45,6 +44,7 @@ class PDFParser:
|
|
|
45
44
|
self._data = data
|
|
46
45
|
|
|
47
46
|
def parse(self) -> List[Page]:
|
|
47
|
+
import deepdoctection as dd
|
|
48
48
|
analyzer = dd.get_dd_analyzer()
|
|
49
49
|
parsed_pages = []
|
|
50
50
|
with tempfile.NamedTemporaryFile(delete=False, suffix=".pdf") as f:
|
|
@@ -68,7 +68,6 @@ class PDFParser:
|
|
|
68
68
|
)
|
|
69
69
|
)
|
|
70
70
|
figures = parsed_page.get_annotation(category_names=dd.LayoutType.FIGURE)
|
|
71
|
-
print(len(figures))
|
|
72
71
|
for figure in figures:
|
|
73
72
|
image_bytes = dd.viz_handler.encode(figure.viz())
|
|
74
73
|
fragments.append(
|
|
@@ -80,7 +79,6 @@ class PDFParser:
|
|
|
80
79
|
)
|
|
81
80
|
|
|
82
81
|
tables = parsed_page.get_annotation(category_names=dd.LayoutType.TABLE)
|
|
83
|
-
print(len(tables))
|
|
84
82
|
for table in tables:
|
|
85
83
|
fragments.append(
|
|
86
84
|
PageFragment(
|
|
@@ -13,13 +13,13 @@ indexify/extractor_sdk/extractor.py,sha256=IEZvr1Qe-dVmTgAeJFAhEyHUW20n4uTEeEass
|
|
|
13
13
|
indexify/extractor_sdk/utils.py,sha256=bW_D2eMWTzcAYZ8Lv7LUKGgOD0cyW77E6gNO3y7iNNA,6234
|
|
14
14
|
indexify/extractors/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
|
15
15
|
indexify/extractors/embedding.py,sha256=Be6X4odSHbkAEm2myxB04RN-Mvb2bFk8uWXxUpY-Z6E,1859
|
|
16
|
-
indexify/extractors/pdf_parser.py,sha256=
|
|
16
|
+
indexify/extractors/pdf_parser.py,sha256=DwHXVbdy-3SDPOo1U_tdp6g5NWs0Tde5h0GoMGTi1WA,2787
|
|
17
17
|
indexify/graph.py,sha256=UdvrpNc-SdD3U27Ee9aTMMYcSOUz__WQWc31oFHV4yQ,3963
|
|
18
18
|
indexify/local_runner.py,sha256=uuMJbnT4qYMSySxsB3lEC7FSjYnJFh5eNZ00zu5gLNw,4387
|
|
19
19
|
indexify/runner.py,sha256=VVmLGF1kAmEuE461Hs0QJFnSvVWtUzYhhQfB1KptYPU,637
|
|
20
20
|
indexify/settings.py,sha256=LSaWZ0ADIVmUv6o6dHWRC3-Ry5uLbCw2sBSg1e_U7UM,99
|
|
21
21
|
indexify/utils.py,sha256=rDN2lrsAs9noJEIjfx6ukmC2SAIyrlUt7QU-kaBjujM,125
|
|
22
|
-
indexify-0.0.
|
|
23
|
-
indexify-0.0.
|
|
24
|
-
indexify-0.0.
|
|
25
|
-
indexify-0.0.
|
|
22
|
+
indexify-0.0.42.dist-info/LICENSE.txt,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
|
|
23
|
+
indexify-0.0.42.dist-info/METADATA,sha256=kA0oqiewM71YLHKk-cXX-nR7ERJDkC9rVZqjF2LX_zI,1913
|
|
24
|
+
indexify-0.0.42.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
|
|
25
|
+
indexify-0.0.42.dist-info/RECORD,,
|
|
File without changes
|
|
File without changes
|