✨ Google Drive Source #31458

flash1293 · 2023-10-16T16:00:50Z

What

This PR adds a new source connector for google drive. It is based on the File CDK and allows for all features provided this way.

Features:

Authentication via OAuth or Service Account JSON
Sync files in specified folder (including subfolders)
Google Docs are read as Docx files
Other Google office formats (slides, sheets, drawing) are exported as PDF

⚠️ This is not using the smart_open library, instead the whole file is downloaded via the MediaIoBaseDownload class provided by the Google SDK. It should be possible to leverage smart_open (PoC is on a separate PR here: #31866), but I would like to split it out of the initial version ⚠️

⚠️ This is currently only doing file type detection via the file name, I opened https://github.com/airbytehq/airbyte-internal-issues/issues/2219 to track to improve this ⚠️

File paths always start at the defined folder.

How to test

OAuth

OAuth is not super simple to to test. The general steps are defined on this notion page: https://www.notion.so/Testing-OAuth-Locally-fea17aeb14c74cacb5f3ed856daae753

Check out https://github.com/airbytehq/airbyte-platform-internal/pull/9359
Go to oss directory cd oss
In the airbyte-webapp project, enable OAuth under oss/airbyte-webapp/src/core/services/features/constants.ts:

const defaultOssFeatures = [
	// ...Leave all the existing features and add...
	FeatureItem.AllowOAuthConnector
];

Re-build platform images ./gradlew assemble
Start platform via BASIC_AUTH_USERNAME="" BASIC_AUTH_PASSWORD="" VERSION=dev docker-compose --file ./docker-compose.yaml up
Build google drive connector airbyte-ci connectors --name=source-google-drive build
Add as custom connector (airbyte/source-google-drive.dev)
Run docker exec -ti airbyte-db psql -U docker -d airbyte -c "select id from actor_definition where name like '%Drive%';" to get the source deifnition id
Get OAuth client id and client secret from password manager (search for "Google Drive OAuth Test App")
Add oauth credentials to local Airbyte instance:

API_URL=localhost:8001/api/v1;

curl -X POST \
   -H 'Content-type: application/json' \
   $API_URL/source_oauths/oauth_params/create \
   -d '{"sourceDefinitionId": "<the id from above>", "params": {"credentials": {"client_id": "___", "client_secret": "___" }}}'

Navigate to localhost:8000
Create a source

Service account json

Get OAuth service account jsonfrom password manager (search for "Google Drive OAuth Test App")
Note that this service account will not have access to your Google Drive files (to test with it the folder needs to be public)

vercel · 2023-10-16T16:00:56Z

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
airbyte-docs	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Nov 2, 2023 0:24am

github-actions · 2023-10-16T16:01:10Z

Before Merging a Connector Pull Request

Wow! What a great pull request you have here! 🎉

To merge this PR, ensure the following has been done/considered for each connector added or updated:

PR name follows PR naming conventions
Breaking changes are considered. If a Breaking Change is being introduced, ensure an Airbyte engineer has created a Breaking Change Plan.
Connector version has been incremented in the Dockerfile and metadata.yaml according to our Semantic Versioning for Connectors guidelines
You've updated the connector's metadata.yaml file any other relevant changes, including a breakingChanges entry for major version bumps. See metadata.yaml docs
Secrets in the connector's spec are annotated with airbyte_secret
All documentation files are up to date. (README.md, bootstrap.md, docs.md, etc...)
Changelog updated in docs/integrations/<source or destination>/<name>.md with an entry for the new version. See changelog example
Migration guide updated in docs/integrations/<source or destination>/<name>-migrations.md with an entry for the new version, if the version is a breaking change. See migration guide example
If set, you've ensured the icon is present in the platform-internal repo. (Docs)

If the checklist is complete, but the CI check is failing,

Check for hidden checklists in your PR description
Toggle the github label checklist-action-run on/off to re-run the checklist CI.

…ogle-drive

…airbyte into flash1293/source-google-drive

airbyte-oss-build-runner · 2023-10-23T16:31:54Z

source-google-drive test report (commit `c54003ad48`) - ❌

⏲️ Total pipeline duration: 01mn00s

Step	Result
Build source-google-drive docker image for platform(s) linux/amd64	✅
Unit tests	❌
Code format checks	❌
Validate metadata for source-google-drive	✅
Connector version semver check	✅
Connector version increment check	✅
QA checks	✅

🔗 View the logs here

☁️ View runs for commit in Dagger Cloud

Please note that tests are only run on PR ready for review. Please set your PR to draft mode to not flood the CI engine and upstream service on following commits.
You can run the same pipeline locally on this branch with the airbyte-ci tool with the following command

airbyte-ci connectors --name=source-google-drive test

airbyte-oss-build-runner · 2023-10-23T16:45:48Z

source-google-drive test report (commit `54157ef2e9`) - ❌

⏲️ Total pipeline duration: 28.81s

Step	Result
Build source-google-drive docker image for platform(s) linux/amd64	✅
Unit tests	❌
Code format checks	❌
Validate metadata for source-google-drive	✅
Connector version semver check	✅
Connector version increment check	✅
QA checks	✅

🔗 View the logs here

☁️ View runs for commit in Dagger Cloud

Please note that tests are only run on PR ready for review. Please set your PR to draft mode to not flood the CI engine and upstream service on following commits.
You can run the same pipeline locally on this branch with the airbyte-ci tool with the following command

airbyte-ci connectors --name=source-google-drive test

…ogle-drive

clnoll

LGTM!

flash1293 · 2023-10-27T14:47:44Z

@aaronsteers Worked through the comments, unstructured files are also supported now. One open question is whether we should allow multiple streams or simplify the configuration object to just point to a folder and specify a glob pattern to fetch. What do you think?

…ogle-drive

flash1293 · 2023-10-30T11:34:38Z

@aaronsteers As discussed, I added a check to avoid fetching prefixes in most cases. It's not perfect, (e.g. for subfolder/*.pdf it will still fetch folders within subfolder even though they have no chance of matching), but that seemed like an edge case that's hard to work around, so I will not cover it for now and open a separate issue. Could you take another look please?

docs/integrations/sources/google-drive.md

aaronsteers

I've added a number of questions, most around UX, paths, and globs.

@flash1293 - We have a 1:1 tomorrow so we can also talk through this if helpful and more efficient to process that way.

docs/integrations/sources/google-drive.md

aaronsteers · 2023-10-30T19:46:23Z

docs/integrations/sources/google-drive.md

+
+One record will be emitted for each document. Keep in mind that large files can emit large records that might not fit into every destination as each destination has different limitations for string fields.
+
+Google documents are exported as Docx files while spreadsheets, presentations and drawings are exported as PDF files.


Probably a nit, but this seemed it could be misinterpreted as us passing a PDF document downstream. A bit tricky to explain concisely, but a possible alternate wording might be something like:

Suggested change

Google documents are exported as Docx files while spreadsheets, presentations and drawings are exported as PDF files.

Before parsing each document, the connector exports Google Document files to Docx format internally. Google Sheets, Google Slides, and drawings are internally exported and parsed by the connector as PDFs.

Or maybe this is simpler:

Suggested change

Google documents are exported as Docx files while spreadsheets, presentations and drawings are exported as PDF files.

Google documents are processed as Docx files while spreadsheets, presentations and drawings are processed as PDFs.

Totally up to you.

I like it much better, thanks

aaronsteers · 2023-10-30T19:48:12Z

docs/integrations/sources/google-drive.md

+## User Schema
+
+Providing a schema allows for more control over the output of this stream. Without a provided schema, columns and datatypes will be inferred from the first created file in the bucket matching your path pattern and suffix. This will probably be fine in most cases but there may be situations you want to enforce a schema instead, e.g.:


Do we need a callout, here or in the unstructured doc section, which marks this custom schema section as not applying to unstructured text docs?

Good point, added a callout

aaronsteers · 2023-10-30T19:57:23Z

airbyte-integrations/connectors/source-google-drive/source_google_drive/stream_reader.py

+    def _get_export_mime_type(self, file: GoogleDriveRemoteFile):
+        """
+        Returns the mime type to export Google App documents as.
+
+        Google Docs are exported as Docx to preserve as much formatting as possible, everything else goes through PDF.
+        """


Wdyt of adding this inline for easy reference?

Suggested change

def _get_export_mime_type(self, file: GoogleDriveRemoteFile):

"""

Returns the mime type to export Google App documents as.

Google Docs are exported as Docx to preserve as much formatting as possible, everything else goes through PDF.

"""

def _get_export_mime_type(self, file: GoogleDriveRemoteFile):

"""

Returns the mime type to export Google App documents as.

Google Docs are exported as Docx to preserve as much formatting as possible,

everything else goes through PDF. Supported export formats are documented here:

- https://developers.google.com/drive/api/guides/ref-export-formats

"""

aaronsteers · 2023-10-30T20:08:25Z

docs/integrations/sources/google-drive.md

+Let's look at a specific example, matching the following bucket layout:
+
+```text
+myBucket


Suggested change

Let's look at a specific example, matching the following bucket layout:

```text

myBucket

Let's look at a specific example, matching the following drive layout:

```text

MyDrive

aaronsteers · 2023-10-30T20:13:33Z

airbyte-integrations/connectors/source-google-drive/source_google_drive/stream_reader.py

+        if file["mimeType"] == GOOGLE_DOC_MIME_TYPE:
+            return f"{file['name']}.docx"
+        elif self._is_exportable_document(file["mimeType"]):
+            return f"{file['name']}.pdf"
+        else:
+            return file["name"]


Is it necessary for us to add the extension? I ask because it is possible for these to create path collisions. Users could have "My File" and "My File.docx" or "My File.pdf" already in the same directory. When dealing with sharing files across Google and non-Google audiences, it does happen to sometimes have the same file twice in a directory, once as Google native format and ones as the more portable non-google format.

In future, we may also want to customize handling for presentations. Similar to the DOCX format, a PPTX format is probably going to have a lower cost parsing for slide decks, but I don't think we need to prioritize it as of yet and there might be some areas where slides are better parsed as PDFs than as slides. 🤷 (Testing would tell.)

A larger issue, I just realized, is Google Sheets. They don't really fit the 'text documents' mold, and it might (?) be preferably to perhaps ignore them when this parser is selected. Presumably later on, this connector could add a "Google Sheets" parser that took the same parse options as Excel and internally converted to Excel before parsing. 🤷

I'm thinking now also about other document types like CSVs - whether they would+should get picked up by the text document parser, or if we should ignore them based on their MIME type when the Document Parser is selected.

I wonder if an efficient way to handle this would be to add a MIME type glob pattern. Traditionally, we'd use file extensions for this, but google docs don't have them, and a more robust include/exclude logic would perhaps necessarily use MIME type or some other means of filtering. (Caveat: this is an area I want to understand well and plan for, but I don't know if it should block merging.)

You are raising a very good point about robustness of the file detection - what about the following:

There are three "leads" we can leverage:

The mime type

The file extension

The actual file (inspecting the first few bytes) - this works well for docx, pptx and pdf so we should be pretty safe here

I created a PR to add this to the existing unstructured parser in the CDK here: #31997

Once that is merged I can remove the artificial file extensions from this PR and pass along the mime type (S3 and Azure Blob Storage also have a notion of mime type so it should be possible to extend them in a similar way, but as we have the file inspection it shouldn't be strictly necessary - plus it's always good to have another layer to fall back to as not all objects are guaranteed to have this information).

A larger issue, I just realized, is Google Sheets. They don't really fit the 'text documents' mold, and it might (?) be preferably to perhaps ignore them when this parser is selected. Presumably later on, this connector could add a "Google Sheets" parser that took the same parse options as Excel and internally converted to Excel before parsing. 🤷

Yeah, I wasn't sure about that either - we already have a dedicated Sheets connector, maybe that's good enough for now. I'm happy to exclude it for the time being and get back to it later.

Are you planning to merge #31997 before this PR is merged? (That seems preferable to me.)

Yes, I would like to take that route.

aaronsteers · 2023-10-30T20:16:10Z

airbyte-integrations/connectors/source-google-drive/source_google_drive/stream_reader.py

+            if mode == FileReadMode.READ:
+                raise ValueError("Cannot read Google Docs/Sheets/Presentations and so on as text. Please set the format to PDF")


Is this an instruction for the developer or the user?

If to the developer, I could see the guidance of "please set the format to PDF" being potentially confusing.

It's for the user, I adjusted the error message to be more clear/actionable

aaronsteers · 2023-10-30T20:18:56Z

airbyte-integrations/connectors/source-google-drive/source_google_drive/stream_reader.py

+EXPORTABLE_DOCUMENTS_MIME_TYPES = [
+    GOOGLE_DOC_MIME_TYPE,
+    "application/vnd.google-apps.spreadsheet",
+    "application/vnd.google-apps.presentation",
+    "application/vnd.google-apps.drawing",
+]


Wdyt of leaving off Google Sheets (application/vnd.google-apps.spreadsheet) for now?

It may be preferable to pass these through the Excel format parser, rather than parsing as text documents.

cc @clnoll

Good point @aaronsteers, we most certainly will want a separate parser for Excel eventually.

I can definitely see a situation where a user doesn't really care what files are in their folder they just want everything that can be parsed. But, if we wanted to offer that I think maybe it would be best done as an umbrella parser that can call out to other parsers rather than including Excel in with this one.

Agree with the points made here, let's just not process them for now (we have the Sheets parser already anyway)

aaronsteers · 2023-10-30T20:27:35Z

airbyte-integrations/connectors/source-google-drive/integration_tests/spec.json

+      "folder_url": {
+        "title": "Folder Url",
+        "description": "URL for the folder you want to sync",
+        "order": 0,
+        "type": "string"
+      },


I am not sure I understand exactly how this global folder URL setting interacts with stream-level globs, and it might be helpful to add specific guidance on this front. In the readme, there's reference to the "bucket" (or "drive") name not needing to be included in the glob pattern, I'm not clear if that would extend to receiving a subfolder here as the base folder URL.

For instance, if folder URL points to the MyDrive/Engineering/Tech Specs folder, would the user provide a glob of ** or MyDrive/Engineering/Tech Specs/** to get all files in that directory?

I think providing just the subfolder glob is probably preferable but either way could work in theory.

The glob is just for the subfolder (the specified global folder URL is like the "bucket"). I will clarify this in the documentation

aaronsteers · 2023-10-30T21:02:14Z

airbyte-integrations/connectors/source-google-drive/source_google_drive/stream_reader.py

+                        remote_file = GoogleDriveRemoteFile(
+                            uri=file_name, last_modified=last_modified, id=new_file["id"], mimeType=new_file["mimeType"]
+                        )
+                        if self.file_matches_globs(remote_file, globs):
+                            yield remote_file


Following also my comments in _get_file_name(), it doesn't feel right to use our appended suffixes (.docx, .pdf) for documents that don't have those suffixes. Could we remove that suffix without breaking any key functionality?

Also, just to confirm, do we process all files in the included directory, except those excluded by a users' glob pattern? Or do we have some kind of internal filter when using the document parser? I'm a bit worried this could break if a non-readable non-text file (.zip, .ico, .parquet) exists in the directory.

As mentioned in the other comment, once #31997 is merged we shouldn't need to append suffixes anymore.

Also, just to confirm, do we process all files in the included directory, except those excluded by a users' glob pattern? Or do we have some kind of internal filter when using the document parser? I'm a bit worried this could break if a non-readable non-text file (.zip, .ico, .parquet) exists in the directory.

As the fact whether a file is readable or not is a property of the parser, not the stream reader, this can't be controlled here. See our previous discussion of this over here: #31209 (comment)

Originally I wanted to go with the more lenient behavior of just skipping the file if it can't be read exactly for the reason you stated - if you just want to point to a folder and have it pick up everything that works somewhat, then skipping is probably better because otherwise it's relatively simple to fail the sync / connection check which could lead to an annoying user experience.

However, skipping is also not perfect - we have no good way to indicate what happened in the UI in this kind of situation, so another problematic behavior could be the user pointing to a folder full of files we can't process, the sync running "successfully" with minimal data actually arriving in the destination.

It comes down to what I highlighted already - it's hard to tell what's right as long as people didn't try to actually use it. In this kind of situation, going with the more strict approach is actually not a bad choice because we can relax requirements later on without disturbing existing users while going the opposite route would be a breaking change. I'm OK with going both ways here though - a third way would be to make it a configuration option of the document file type parser ("ignore unparsable files")

A takeaway from this I would like to follow a bit is to extend the protocol with something between a failed and a successful sync (something like "succeeded with warnings" or whatever) that could be used in this situation (the warning could be "encountered 1234 files that couldn't be processed")

I like the idea of a failure with warnings. That seems like the best of both worlds to me.

…ogle-drive

flash1293 · 2023-10-31T12:50:14Z

@clnoll @aaronsteers Thanks for the review. I made the following changes:

Improved docs
Improved file type detection logic in CDK in separate PR
Removed handling of Google Sheets files

The biggest open question I'm seeing is whether unprocessable files should be ignored or fail the sync or whether this should be configurable by the user.

flash1293 · 2023-11-01T10:54:03Z

@aaronsteers My preferred option would be to publish as is (after the improved file format detection is added) and adding a user option to configure fail vs. skip in the document file format parser in a follow-up PR after the initial release.

aaronsteers

Thanks for all of the changes!

I'm excited to see this ship, and I don't think we need to block for the handling of unhandled file types.

Ideally, I think we'd have:

The parser selection dictates which extensions and/or MIME types are eligible to be processed.
Qualified files which cannot be processed would trigger an error. (E.g. a corrupted PDF would still fail the Text Document parser, since its eligible for parsing and can't be parsed.)
Files which are ineligible for the selected file parser would not be attempted to be processed. For visibility, we can optionally log "skipped" files in connector logs.
Configurability:
- Config option, disabled by default, to try to parse everything, even ineligible files. (Basically an "opt-in" to fail on non-eligible files existing, but we try anyway.)
- MIME Type Filers. Either a glob, set of globs, or list of MIME types to include/exclude.
- Glob filtering. (No change. This already exists, and can be used to filter on file name extensions.)

Above is non-blocking. The above, or some variation or alteration of these, could be implemented in one or more subsequent iterations.

aaronsteers · 2023-11-02T07:10:31Z

@aaronsteers My preferred option would be to publish as is (after the improved file format detection is added) and adding a user option to configure fail vs. skip in the document file format parser in a follow-up PR after the initial release.

Sounds great! Added comment above with some ideas on potential configurability.

…ogle-drive

LuongNV1 · 2023-12-04T04:14:23Z

I want authenticate via google with only button

flash1293 · 2023-12-04T15:21:10Z

@LuongNV1 This is currently only possible via Airbyte Cloud

wip

5313a65

octavia-squidington-iii added the connectors/source/google-drive label Oct 16, 2023

Automated Commit - Formatting Changes

d480cd1

octavia-squidington-iii added the area/connectors Connector related issues label Oct 16, 2023

Joe Reuter and others added 12 commits October 18, 2023 17:39

Merge remote-tracking branch 'origin/master' into flash1293/source-go…

450b24e

…ogle-drive

start with oauth

83717fe

Automated Commit - Formatting Changes

84cbbde

make oauth work

688f74e

handle google docs and loops in the folder structure

044816d

Automated Commit - Formatting Changes

a417345

clean up

d55c445

Automated Commit - Formatting Changes

c3787c1

fixes and tests

ddc0837

Merge remote-tracking branch 'origin/master' into flash1293/source-go…

63d61a4

…ogle-drive

Merge branch 'flash1293/source-google-drive' of github.com:airbytehq/…

ea938aa

…airbyte into flash1293/source-google-drive

start to add tests

9c6fb0a

octavia-squidington-iii added the area/documentation Improvements or additions to documentation label Oct 23, 2023

vercel bot deployed to Preview October 23, 2023 15:44 View deployment

work on tests

c54003a

flash1293 marked this pull request as ready for review October 23, 2023 16:28

octavia-squidington-iv requested a review from a team October 23, 2023 16:30

Automated Commit - Formatting Changes

54157ef

Joe Reuter added 4 commits October 25, 2023 13:36

stuff

893ee5c

create tests

890d643

Merge remote-tracking branch 'origin/master' into flash1293/source-go…

9304b4a

…ogle-drive

format

3ac08cc

review comments

0ffa50a

clnoll approved these changes Oct 27, 2023

View reviewed changes

Joe Reuter added 3 commits October 30, 2023 12:25

respect prefixes

3f6a8ea

fix stuff

0a48536

Merge remote-tracking branch 'origin/master' into flash1293/source-go…

3bbbc86

…ogle-drive

vercel bot deployed to Preview October 30, 2023 11:37 View deployment

add pptx support

2918058

vercel bot deployed to Preview October 30, 2023 14:30 View deployment

aaronsteers reviewed Oct 30, 2023

View reviewed changes

docs/integrations/sources/google-drive.md Outdated Show resolved Hide resolved

aaronsteers reviewed Oct 30, 2023

View reviewed changes

Joe Reuter added 2 commits October 31, 2023 13:07

Merge remote-tracking branch 'origin/master' into flash1293/source-go…

018a740

…ogle-drive

review comments

f399b66

flash1293 mentioned this pull request Oct 31, 2023

File CDK unstructured parser: Improve file type detection #31997

Merged

vercel bot deployed to Preview October 31, 2023 12:46 View deployment

octavia-squidington-iv requested a review from a team November 1, 2023 09:33

aaronsteers approved these changes Nov 2, 2023

View reviewed changes

Joe Reuter added 4 commits November 2, 2023 12:48

format

30cbfaa

Merge remote-tracking branch 'origin/master' into flash1293/source-go…

1cd4c2c

…ogle-drive

fix tests

fe9b960

improve doc

838c3fe

vercel bot deployed to Preview November 2, 2023 12:24 View deployment

flash1293 merged commit f487e15 into master Nov 2, 2023
17 checks passed

flash1293 deleted the flash1293/source-google-drive branch November 2, 2023 13:28


		One record will be emitted for each document. Keep in mind that large files can emit large records that might not fit into every destination as each destination has different limitations for string fields.

		Google documents are exported as Docx files while spreadsheets, presentations and drawings are exported as PDF files.

	Google documents are exported as Docx files while spreadsheets, presentations and drawings are exported as PDF files.
	Before parsing each document, the connector exports Google Document files to Docx format internally. Google Sheets, Google Slides, and drawings are internally exported and parsed by the connector as PDFs.

	Google documents are exported as Docx files while spreadsheets, presentations and drawings are exported as PDF files.
	Google documents are processed as Docx files while spreadsheets, presentations and drawings are processed as PDFs.

		## User Schema

		Providing a schema allows for more control over the output of this stream. Without a provided schema, columns and datatypes will be inferred from the first created file in the bucket matching your path pattern and suffix. This will probably be fine in most cases but there may be situations you want to enforce a schema instead, e.g.:

		if mode == FileReadMode.READ:
		raise ValueError("Cannot read Google Docs/Sheets/Presentations and so on as text. Please set the format to PDF")

✨ Google Drive Source #31458

✨ Google Drive Source #31458

Conversation

flash1293 commented Oct 16, 2023 • edited Loading

What

How to test

OAuth

Service account json

vercel bot commented Oct 16, 2023 • edited Loading

github-actions bot commented Oct 16, 2023 • edited by flash1293 Loading

Before Merging a Connector Pull Request

airbyte-oss-build-runner commented Oct 23, 2023

source-google-drive test report (commit c54003ad48) - ❌

airbyte-oss-build-runner commented Oct 23, 2023

source-google-drive test report (commit 54157ef2e9) - ❌

clnoll left a comment

Choose a reason for hiding this comment

flash1293 commented Oct 27, 2023

flash1293 commented Oct 30, 2023

aaronsteers left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

aaronsteers Oct 30, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

aaronsteers Oct 30, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

flash1293 Oct 31, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

flash1293 commented Oct 31, 2023

flash1293 commented Nov 1, 2023

aaronsteers left a comment • edited Loading

Choose a reason for hiding this comment

aaronsteers commented Nov 2, 2023

LuongNV1 commented Dec 4, 2023

flash1293 commented Dec 4, 2023

flash1293 commented Oct 16, 2023 •

edited

Loading

vercel bot commented Oct 16, 2023 •

edited

Loading

github-actions bot commented Oct 16, 2023 •

edited by flash1293

Loading

source-google-drive test report (commit `c54003ad48`) - ❌

source-google-drive test report (commit `54157ef2e9`) - ❌

aaronsteers Oct 30, 2023 •

edited

Loading

aaronsteers Oct 30, 2023 •

edited

Loading

flash1293 Oct 31, 2023 •

edited

Loading

aaronsteers left a comment •

edited

Loading