pengjunjie/ragflow
1
0
Fork 1
mirror of https://github.com/infiniflow/ragflow.git synced 2025-12-23 06:46:40 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix errors (#11804)

Browse Source 
### What problem does this PR solve?

1. typos
2. grammar errors.

### Type of change

- [x] Refactoring

Signed-off-by: Jin Hai <haijin.chn@gmail.com>
This commit is contained in:
Jin Hai
2025-12-08 12:21:18 +08:00
committed by GitHub
parent 5a2011e687
commit 43f51baa96
25 changed files with 53 additions and 54 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
common/data_source/google_drive/doc_conversion.py
View File

@ -76,7 +76,7 @@ ALL_ACCEPTED_FILE_EXTENSIONS = ACCEPTED_PLAIN_TEXT_FILE_EXTENSIONS + ACCEPTED_DO
MAX_RETRIEVER_EMAILS = 20
CHUNK_SIZE_BUFFER = 64 # extra bytes past the limit to read
# This is not a standard valid unicode char, it is used by the docs advanced API to
# This is not a standard valid Unicode char, it is used by the docs advanced API to
# represent smart chips (elements like dates and doc links).
SMART_CHIP_CHAR = "\ue907"
WEB_VIEW_LINK_KEY = "webViewLink"

4
common/data_source/google_drive/file_retrieval.py
View File

@ -141,7 +141,7 @@ def crawl_folders_for_files(
# Only mark a folder as done if it was fully traversed without errors
# This usually indicates that the owner of the folder was impersonated.
# In cases where this never happens, most likely the folder owner is
# not part of the google workspace in question (or for oauth, the authenticated
# not part of the Google Workspace in question (or for oauth, the authenticated
# user doesn't own the folder)
if found_files:
update_traversed_ids_func(parent_id)
@ -232,7 +232,7 @@ def get_files_in_shared_drive(
**kwargs,
):
# If we found any files, mark this drive as traversed. When a user has access to a drive,
# they have access to all the files in the drive. Also not a huge deal if we re-traverse
# they have access to all the files in the drive. Also, not a huge deal if we re-traverse
# empty drives.
# NOTE: ^^ the above is not actually true due to folder restrictions:
# https://support.google.com/a/users/answer/12380484?hl=en

4
common/data_source/google_drive/model.py
View File

@ -22,7 +22,7 @@ class GDriveMimeType(str, Enum):
MARKDOWN = "text/markdown"
# These correspond to The major stages of retrieval for google drive.
# These correspond to The major stages of retrieval for Google Drive.
# The stages for the oauth flow are:
# get_all_files_for_oauth(),
# get_all_drive_ids(),
@ -117,7 +117,7 @@ class GoogleDriveCheckpoint(ConnectorCheckpoint):
class RetrievedDriveFile(BaseModel):
"""
Describes a file that has been retrieved from google drive.
Describes a file that has been retrieved from Google Drive.
user_email is the email of the user that the file was retrieved
by impersonating. If an error worthy of being reported is encountered,
error should be set and later propagated as a ConnectorFailure.