Fix xml name without extrension (#75)

* Update dependencies * Run formatter * Fix issue when pdf files without extension
huridocs · May 9, 2023 · 72e6869 · 72e6869
1 parent 7499717
commit 72e6869
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 2 deletions.
diff --git a/src/app.py b/src/app.py
@@ -123,7 +123,7 @@ async def get_paragraphs(tenant: str, pdf_file_name: str):
 @app.get("/get_xml/{tenant}/{pdf_file_name}", response_class=PlainTextResponse)
 async def get_xml(tenant: str, pdf_file_name: str):
     try:
-        xml_file_name = ".".join(pdf_file_name.split(".")[:-1]) + ".xml"
+        xml_file_name = pdf_file_name.replace(".", "") + ".xml"
 
         with open(
             f"{config.DATA_PATH}/xml/{tenant}/{xml_file_name}",

diff --git a/src/extract_pdf_paragraphs/extract_paragraphs.py b/src/extract_pdf_paragraphs/extract_paragraphs.py
@@ -15,7 +15,7 @@
 
 
 def get_paths(tenant: str, pdf_file_name: str):
-    file_name = "".join(pdf_file_name.split(".")[:-1])
+    file_name = pdf_file_name.replace(".", "")
     pdf_file_path = f"{config.DATA_PATH}/to_extract/{tenant}/{pdf_file_name}"
     xml_file_path = f"{config.DATA_PATH}/xml/{tenant}/{file_name}.xml"
     failed_pdf_path = f"{config.DATA_PATH}/failed_pdf/{tenant}/{pdf_file_name}"