Add Check for empty parts to OpenXmlValidator

[mikeebowen]

• closes Is it possible to check if a file is a valid document and not some other file type renamed to .docx or some other office extension? #816

The issue was not that there was no way to check for a minimally valid document, the issue was that the validator considered completely empty parts to be valid. So this issue was resolved by adding a check for empty parts to the OpenXmlValidator.

[github-actions]

Test Results

    58 files   -   1      58 suites   - 1   54m 55s ⏱️ + 3m 5s
 2 060 tests +  1   2 057 ✅ +  1   3 💤 ±0  0 ❌ ±0 
32 334 runs   - 164  32 298 ✅  - 164  36 💤 ±0  0 ❌ ±0 

Results for commit 31aec5a. ± Comparison against base commit b1d49f3.

♻️ This comment has been updated with latest results.

[Copilot]

Pull Request Overview

This PR introduces the IsMinimumDocument method to WordprocessingDocument, SpreadsheetDocument, and PresentationDocument to validate that a file meets minimal structural and extension requirements before further processing.

• Added IsMinimumDocument method for WordprocessingDocument with path, extension, and Body element checks
• Added IsMinimumDocument method for SpreadsheetDocument with validations for sheets and sheet data and handling for unsupported document types
• Added IsMinimumDocument method for PresentationDocument with validation for NotesSize element and restrictions on unsupported document types

Reviewed Changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 3 comments.

File	Description
src/DocumentFormat.OpenXml/Packaging/WordprocessingDocument.cs	Added validation method ensuring file existence, correct extension, and Body element presence
src/DocumentFormat.OpenXml/Packaging/SpreadsheetDocument.cs	Introduced method validation for SpreadsheetDocument with extension checks and element validations
src/DocumentFormat.OpenXml/Packaging/PresentationDocument.cs	Added a validation method ensuring PresentationDocument contains valid NotesSize and proper extension

[twsouthwick]

This is a great idea. However, it would be great for it to have the following characteristics:

• Be overrideable/replaceable
• Be accessed off of OpenXmlPackage directly rather than implemented separately

Questions that would decide the API shape:

• Should this be a separate API? Or should it be opt in while opening a package? I tend not to be a fan of APIs that check if something is valid then doing it - there's always the potential something gets changed underneath
• Do we really want to validate the extension? What about if it's a stream/package? If it is overrideable, we could have a check if we've opened it with a path to validate that if needed
• Are we OK with it throwing an exception if it is an ill-formed package (i.e. we already do that)?

[mikeebowen]

@twsouthwick

Be overrideable/replaceable

Created virtual method in OpenXmlPackage

Be accessed off of OpenXmlPackage directly rather than implemented separately

There is a default implementation in OpenXmlPackage but the individual Wordprocessing, Spreadsheet, and Presentation Documents have unique implementations for their different requirements.

Should this be a separate API? Or should it be opt in while opening a package? I tend not to be a fan of APIs that check if something is valid then doing it - there's always the potential something gets changed underneath

Added VerifyMinimumPackage option to OpenSettings when true, the package contents are examined and throw an exception if minimum package requirements are not met

Do we really want to validate the extension? What about if it's a stream/package? If it is overrideable, we could have a check if we've opened it with a path to validate that if needed

The minimum package is verified on all Open overloads with VerifyMinimumPackage true. The Open method handles exceptions for the file path such as invalid extension or invalid characters, so with the new API validating the path is not necessary.

Are we OK with it throwing an exception if it is an ill-formed package (i.e. we already do that)?

Removed file checks. The Open method already handles those.

[twsouthwick]

@mikeebowen this would be a good use case for a feature so it can be overridden at runtime vs compile time. Here's an example of how to do it: fa893dd

[twsouthwick]

A few questions about the direction/goal here

Only a few kinds of packages are supported (such as in the presentation doc) - it will fail if it's an unsupported one

Is this what we want? I'd expect it to validate it but fail if it is invalid. It appears to also fail if we don't have a check which seems a little unexpected.

Validate() throws and returns true/false

I'd expect it to do one or the other. Is the manual throwing due to custom messages? If so, let's make it consistent

Mixture of exception types

I think we should use an existing OpenXmlPackageException type or create a new one for this

How to think about this vs validation

At some point, how is this different from the existing OpenXmlValidator logic? Should we be hooking this up to that and performing a full validation?

[twsouthwick]

need to dispose the stream from getpart. probably best to add an extension method of ".IsEmptyPart()" that can handle correctly disposing the stream