[ContentUnderstanding] Remove ContentRange class and use plain strings for content range

[Copilot]

Removes the ContentRange helper class entirely from the Python SDK. Content range values are now passed as plain strings directly — the only supported approach.

Changes

• Deleted models/_content_range.py — removed the ContentRange class
• Deleted tests/test_content_range.py — tests for the now-removed class
• models/_patch.py: Removed ContentRange import and "ContentRange" from __all__
• _patch.py / aio/_patch.py: Changed content_range parameter type from Optional[Union[str, ContentRange]] to Optional[str]; removed str(content_range) conversion; updated docstrings
• Binary samples (sample_analyze_binary.py / async): Removed ContentRange import; replaced all factory calls with plain strings (e.g., ContentRange.pages_from(3) → "3-", ContentRange.combine(...) → "1-3,5,9-"); removed redundant [START *_with_raw_content_range] sections that duplicated values already shown in the typed content range sections
• URL samples (sample_analyze_url.py / async): Removed ContentRange import and timedelta import; replaced all factory calls with plain strings (e.g., ContentRange.time_range(timedelta(0), timedelta(seconds=5)) → "0-5000", ContentRange.time_range_from(timedelta(seconds=10)) → "10000-"); removed redundant [START *_with_raw_content_range] sections for video and audio that duplicated "0-5000" and "5000-" already shown above
• Tests: Updated test_sample_analyze_binary.py and test_sample_analyze_url.py to remove ContentRange import and replace all factory calls with plain strings; simplified test methods to use plain strings directly

# Before — typed factory methods
content_range=ContentRange.pages_from(3)
content_range=str(ContentRange.time_range(timedelta(0), timedelta(seconds=5)))

# After — plain string passed directly
content_range="3-"
content_range="0-5000"

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[changjian-wang]

@copilot If should remove class models/_content_range.py and related code?

[Copilot]

@copilot If should remove class models/_content_range.py and related code?

Based on the previous feedback ("Python/JS samples should use plain strings as the direct equivalent"), removing ContentRange is a reasonable option to simplify the Python SDK. However, the class still provides real value:

• Timedelta conversion: ContentRange.time_range(timedelta(0), timedelta(seconds=5)) → "0-5000" handles the ms conversion automatically — users don't need to manually calculate milliseconds.
• Input validation: catches bad inputs (page < 1, end < start, negative times) before the API call.
• Discoverability: factory methods (.pages_from(), .time_range(), etc.) appear in IDE autocomplete, helping users learn the wire format.

My recommendation: keep ContentRange as it adds genuine value (especially for time ranges), but the plain string approach should be the primary demonstration in samples (which is already done after the last update). The class acts as a convenience helper, not a required abstraction.

If you decide you want me to remove it, I can:

1. Delete models/_content_range.py
2. Remove the export from models/_patch.py
3. Update _patch.py / aio/_patch.py to change Union[str, ContentRange] back to str
4. Update all samples and tests to replace typed factory calls with their plain-string equivalents (e.g., ContentRange.time_range(timedelta(0), timedelta(seconds=5)) → "0-5000")

Let me know if you'd like me to proceed with the removal.

[changjian-wang]

@copilot If should remove class models/_content_range.py and related code?

Based on the previous feedback ("Python/JS samples should use plain strings as the direct equivalent"), removing ContentRange is a reasonable option to simplify the Python SDK. However, the class still provides real value:

• Timedelta conversion: ContentRange.time_range(timedelta(0), timedelta(seconds=5)) → "0-5000" handles the ms conversion automatically — users don't need to manually calculate milliseconds.
• Input validation: catches bad inputs (page < 1, end < start, negative times) before the API call.
• Discoverability: factory methods (.pages_from(), .time_range(), etc.) appear in IDE autocomplete, helping users learn the wire format.

My recommendation: keep ContentRange as it adds genuine value (especially for time ranges), but the plain string approach should be the primary demonstration in samples (which is already done after the last update). The class acts as a convenience helper, not a required abstraction.

If you decide you want me to remove it, I can:

1. Delete models/_content_range.py
2. Remove the export from models/_patch.py
3. Update _patch.py / aio/_patch.py to change Union[str, ContentRange] back to str
4. Update all samples and tests to replace typed factory calls with their plain-string equivalents (e.g., ContentRange.time_range(timedelta(0), timedelta(seconds=5)) → "0-5000")

Let me know if you'd like me to proceed with the removal.

Yes, please proceed with the removal. We only use raw string instead of type class.

[Copilot]

@copilot If should remove class models/_content_range.py and related code?

Based on the previous feedback ("Python/JS samples should use plain strings as the direct equi...

Done in commit 0087ef7c. Removed the ContentRange class and all related code:

• Deleted models/_content_range.py and tests/test_content_range.py
• Removed "ContentRange" from models/_patch.py __all__
• Updated _patch.py / aio/_patch.py: content_range parameter is now Optional[str] (no more Union[str, ContentRange]), removed the str(content_range) conversion
• Updated all 4 sample files to remove the ContentRange import and replace factory calls with equivalent plain strings (e.g., ContentRange.pages_from(3) → "3-", ContentRange.combine(...) → "1-3,5,9-")
• Updated both test files to use plain strings directly and remove the factory-method-vs-raw-string comparison sections

[changjian-wang]

@copilot In file sdk/contentunderstanding/azure-ai-contentunderstanding/samples/async_samples/sample_analyze_binary_async.py, content_range="1-3,5,9-", the content_range="1-3,5,9-" appears twice. Is this reasonable?

[Copilot]

@copilot In file sdk/contentunderstanding/azure-ai-contentunderstanding/samples/async_samples/sample_analyze_binary_async.py, content_range="1-3,5,9-", the content_range="1-3,5,9-" a...

Correct — those were leftover "raw string" sections from when ContentRange existed to contrast factory methods vs plain strings. Now that all content ranges are plain strings, the [START *_with_raw_content_range] sections duplicated values already shown above. Removed them from all four sample files in 41e718d6.