Spaces:
Running
Running
Sakshi
commited on
Commit
·
b099df3
1
Parent(s):
0557bd1
changed ContentFormat to DocumentContentFormat in ocr.py
Browse files- policy_analyser/ocr.py +2 -2
policy_analyser/ocr.py
CHANGED
@@ -7,7 +7,7 @@
|
|
7 |
import pymupdf4llm, pymupdf
|
8 |
from azure.core.credentials import AzureKeyCredential
|
9 |
from azure.ai.documentintelligence import DocumentIntelligenceClient
|
10 |
-
from azure.ai.documentintelligence.models import AnalyzeDocumentRequest,
|
11 |
|
12 |
from policy_analyser import AZURE_LAYOUT_ENDPOINT, AZURE_LAYOUT_KEY, AZURE_LAYOUT_MODEL
|
13 |
|
@@ -44,7 +44,7 @@ class AzureDocumentIntelligenceOCR:
|
|
44 |
poller = self.client.begin_analyze_document(
|
45 |
AZURE_LAYOUT_MODEL,
|
46 |
AnalyzeDocumentRequest(bytes_source = file_bytes),
|
47 |
-
output_content_format =
|
48 |
)
|
49 |
result = poller.result()
|
50 |
return result.content, None
|
|
|
7 |
import pymupdf4llm, pymupdf
|
8 |
from azure.core.credentials import AzureKeyCredential
|
9 |
from azure.ai.documentintelligence import DocumentIntelligenceClient
|
10 |
+
from azure.ai.documentintelligence.models import AnalyzeDocumentRequest, DocumentContentFormat, AnalyzeResult
|
11 |
|
12 |
from policy_analyser import AZURE_LAYOUT_ENDPOINT, AZURE_LAYOUT_KEY, AZURE_LAYOUT_MODEL
|
13 |
|
|
|
44 |
poller = self.client.begin_analyze_document(
|
45 |
AZURE_LAYOUT_MODEL,
|
46 |
AnalyzeDocumentRequest(bytes_source = file_bytes),
|
47 |
+
output_content_format = DocumentContentFormat.MARKDOWN
|
48 |
)
|
49 |
result = poller.result()
|
50 |
return result.content, None
|