From b68edb36c3c255f99444db97c973f2fcd492015d Mon Sep 17 00:00:00 2001
From: Kevin Hu <kevinhu.sh@gmail.com>
Date: Wed, 18 Sep 2024 19:13:37 +0800
Subject: [PATCH] fix parser for pptx of which files are from filemanager
 (#2482)

### What problem does this PR solve?

#2467

### Type of change

- [x] Bug Fix (non-breaking change which fixes an issue)
---
 api/apps/file2document_app.py   |  3 +--
 api/db/services/file_service.py | 18 +++++++++---------
 2 files changed, 10 insertions(+), 11 deletions(-)

diff --git a/api/apps/file2document_app.py b/api/apps/file2document_app.py
index 124f9c57e0..1e4b2c9ad3 100644
--- a/api/apps/file2document_app.py
+++ b/api/apps/file2document_app.py
@@ -77,7 +77,7 @@ def convert():
                     doc = DocumentService.insert({
                         "id": get_uuid(),
                         "kb_id": kb.id,
-                        "parser_id": kb.parser_id,
+                        "parser_id": FileService.get_parser(file.type, file.name, kb.parser_id),
                         "parser_config": kb.parser_config,
                         "created_by": current_user.id,
                         "type": file.type,
@@ -85,7 +85,6 @@ def convert():
                         "location": file.location,
                         "size": file.size
                     })
-                    FileService.set_constant_parser(doc, file.name)
                     file2document = File2DocumentService.insert({
                         "id": get_uuid(),
                         "file_id": id,
diff --git a/api/db/services/file_service.py b/api/db/services/file_service.py
index a81f0c6d7f..48796f6b38 100644
--- a/api/db/services/file_service.py
+++ b/api/db/services/file_service.py
@@ -357,7 +357,7 @@ def upload_document(self, kb, file_objs, user_id):
                 doc = {
                     "id": get_uuid(),
                     "kb_id": kb.id,
-                    "parser_id": kb.parser_id,
+                    "parser_id": self.get_parser(filetype, filename, kb.parser_id),
                     "parser_config": kb.parser_config,
                     "created_by": user_id,
                     "type": filetype,
@@ -366,7 +366,6 @@ def upload_document(self, kb, file_objs, user_id):
                     "size": len(blob),
                     "thumbnail": thumbnail(filename, blob)
                 }
-                self.set_constant_parser(doc, filename)
                 DocumentService.insert(doc)
 
                 FileService.add_file_from_kb(doc, kb_folder["id"], kb.tenant_id)
@@ -377,12 +376,13 @@ def upload_document(self, kb, file_objs, user_id):
         return err, files
 
     @staticmethod
-    def set_constant_parser(doc, filename):
-        if doc["type"] == FileType.VISUAL:
-            doc["parser_id"] = ParserType.PICTURE.value
-        if doc["type"] == FileType.AURAL:
-            doc["parser_id"] = ParserType.AUDIO.value
+    def get_parser(doc_type, filename, default):
+        if doc_type == FileType.VISUAL:
+            return ParserType.PICTURE.value
+        if doc_type == FileType.AURAL:
+            return ParserType.AUDIO.value
         if re.search(r"\.(ppt|pptx|pages)$", filename):
-            doc["parser_id"] = ParserType.PRESENTATION.value
+            return ParserType.PRESENTATION.value
         if re.search(r"\.(eml)$", filename):
-            doc["parser_id"] = ParserType.EMAIL.value
\ No newline at end of file
+            return ParserType.EMAIL.value
+        return default
\ No newline at end of file