From e7dd487779795c9ec9c01585264df2a2bb3d6f25 Mon Sep 17 00:00:00 2001
From: Kevin Hu <kevinhu.sh@gmail.com>
Date: Wed, 18 Sep 2024 09:22:14 +0800
Subject: [PATCH] fix ppt file from filemanager error (#2470)

### 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   |  1 +
 api/db/services/file_service.py | 22 +++++++++++++---------
 2 files changed, 14 insertions(+), 9 deletions(-)

diff --git a/api/apps/file2document_app.py b/api/apps/file2document_app.py
index 0e7301bac..124f9c57e 100644
--- a/api/apps/file2document_app.py
+++ b/api/apps/file2document_app.py
@@ -85,6 +85,7 @@ 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 3e8a80420..a81f0c6d7 100644
--- a/api/db/services/file_service.py
+++ b/api/db/services/file_service.py
@@ -366,14 +366,7 @@ class FileService(CommonService):
                     "size": len(blob),
                     "thumbnail": thumbnail(filename, blob)
                 }
-                if doc["type"] == FileType.VISUAL:
-                    doc["parser_id"] = ParserType.PICTURE.value
-                if doc["type"] == FileType.AURAL:
-                    doc["parser_id"] = ParserType.AUDIO.value
-                if re.search(r"\.(ppt|pptx|pages)$", filename):
-                    doc["parser_id"] = ParserType.PRESENTATION.value
-                if re.search(r"\.(eml)$", filename):
-                    doc["parser_id"] = ParserType.EMAIL.value
+                self.set_constant_parser(doc, filename)
                 DocumentService.insert(doc)
 
                 FileService.add_file_from_kb(doc, kb_folder["id"], kb.tenant_id)
@@ -381,4 +374,15 @@ class FileService(CommonService):
             except Exception as e:
                 err.append(file.filename + ": " + str(e))
 
-        return err, files
\ No newline at end of file
+        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
+        if re.search(r"\.(ppt|pptx|pages)$", filename):
+            doc["parser_id"] = ParserType.PRESENTATION.value
+        if re.search(r"\.(eml)$", filename):
+            doc["parser_id"] = ParserType.EMAIL.value
\ No newline at end of file