DAG Processing Processor attempts to parse PPTX (and other non-code based ZIP archives) based on zipfile.is_zipfile design

[mckenzie-andrew]

Apache Airflow version

Other Airflow 2 version (please specify below)

If "Other Airflow 2 version" selected, which one?

2.10.3

What happened?

Dags Processing Manager attempts to parse PPTX files causing it to place corrupted data in the metadata database.

What you think should happen instead?

Ideally it should not be parsing any none-Python files from the /usr/local/airflow/dags directory.

How to reproduce

Copy text from an Airflow DAG code and place it inside a PPTX slide. Upload the PPTX to /usr/local/airflow/dags, wait for errors to appear in dags processing manager logs (may take little bit of time), for reference I have attached what I used and screenshot:

Testing AIrflow Bug.pptx

Operating System

Amazon Linux 2023

Versions of Apache Airflow Providers

Connection type	Package
AWS Connection	apache-airflow-providers-amazon[aiobotocore]==9.0.0
Postgres Connection	apache-airflow-providers-postgres==5.13.1
FTP Connection	apache-airflow-providers-ftp==3.11.1
Fab Connection	apache-airflow-providers-fab==1.5.0
Celery Connection	apache-airflow-providers-celery==3.8.3
HTTP Connection	apache-airflow-providers-http==4.13.2
IMAP Connection	apache-airflow-providers-imap==3.7.0
Common SQL	apache-airflow-providers-common-sql==1.19.0
SQLite Connection	apache-airflow-providers-sqlite==3.9.0
SMTP Connection	apache-airflow-providers-smtp==1.8.0

Deployment

Amazon (AWS) MWAA

Deployment details

Nothing special, default MWAA deployment no additional configurations or requirements.

Anything else?

Based on how zipfile works --> https://github.com/apache/airflow/blob/main/airflow/utils/file.py#L276

It appears the list of items that could be parsed extends beyond PPTX [1]

Hex signature	ISO 8859-1	Offset	Extension	Description
50 4B 03 04, 50 4B 05 06 (empty archive), 50 4B 07 08 (spanned archive)	PK␃␄,PK␅␆,PK␇␈	0	zip, aar, apk, docx, epub, ipa, jar, kmz, maff, msix, odp, ods, odt, pk3, pk4, pptx, usdz, vsdx, xlsx, xpi	zip file format and formats based on it, such as EPUB, JAR, ODF, OOXML

Are you willing to submit PR?

• Yes I am willing to submit a PR!

Code of Conduct

• I agree to follow this project's Code of Conduct

[boring-cyborg]

Thanks for opening your first issue here! Be sure to follow the issue template! If you are willing to raise PR to address this issue please do so, no need to wait for approval.

[potiuk]

No. The issue is badly diagnosed. Seems like the issue is that spaces are not allowed in preparing the metrics, it has nothing to do with parsing .pptx as zipfile. Can you please create a new issue with better description of this issue, please?

Closing that one as invalid. Ideally also sunmiting a fix with sanitizing the stat name could be done even without raising an issue, so you are most welcome to submit such a fix.