chore(llmobs): dac strip io from anthropic

[jsimpher]

Remove potentially sensitive i/o data from apm spans. This way, prompt and completion data will only appear on the llm obs spans, which are/will be subject to data access controls.

Mostly, this just removes io tag sets. A few things (mostly metrics) have llmobs tags dependent on span tags, so there is a bit more refactoring there.

Let me know if I removed anything that should really stay, or if I missed something that should be restricted.

Checklist

• PR author has checked that all the criteria below are met
• The PR description includes an overview of the change
• The PR description articulates the motivation for the change
• The change includes tests OR the PR description describes a testing strategy
• The PR description notes risks associated with the change, if any
• Newly-added code is easy to change
• The change follows the library release note guidelines
• The change includes or references documentation updates if necessary
• Backport labels are set (if applicable)

Reviewer Checklist

• Reviewer has checked that all the criteria below are met
• Title is accurate
• All changes are related to the pull request's stated goal
• Avoids breaking API changes
• Testing strategy adequately addresses listed risks
• Newly-added code is easy to change
• Release note makes sense to a user of the library
• If necessary, author has acknowledged and discussed the performance implications of this PR as reported in the benchmarks PR comment
• Backport labels are set in a manner that is consistent with the release branch maintenance policy

[github-actions]

CODEOWNERS have been resolved as:

releasenotes/notes/remove-io-data-from-span-anthropic-integration-a8b9ccdd3ac1d22d.yaml  @DataDog/apm-python
ddtrace/contrib/internal/anthropic/_streaming.py                        @DataDog/ml-observability
ddtrace/contrib/internal/anthropic/patch.py                             @DataDog/ml-observability
ddtrace/contrib/internal/anthropic/utils.py                             @DataDog/ml-observability
ddtrace/llmobs/_integrations/anthropic.py                               @DataDog/ml-observability
tests/contrib/anthropic/test_anthropic.py                               @DataDog/ml-observability
tests/snapshots/tests.contrib.anthropic.test_anthropic.test_anthropic_llm.json  @DataDog/ml-observability
tests/snapshots/tests.contrib.anthropic.test_anthropic.test_anthropic_llm_create_image.json  @DataDog/ml-observability
tests/snapshots/tests.contrib.anthropic.test_anthropic.test_anthropic_llm_error.json  @DataDog/ml-observability
tests/snapshots/tests.contrib.anthropic.test_anthropic.test_anthropic_llm_multiple_prompts.json  @DataDog/ml-observability
tests/snapshots/tests.contrib.anthropic.test_anthropic.test_anthropic_llm_multiple_prompts_with_chat_history.json  @DataDog/ml-observability
tests/snapshots/tests.contrib.anthropic.test_anthropic.test_anthropic_llm_stream.json  @DataDog/ml-observability
tests/snapshots/tests.contrib.anthropic.test_anthropic.test_anthropic_llm_stream_helper.json  @DataDog/ml-observability
tests/snapshots/tests.contrib.anthropic.test_anthropic.test_anthropic_llm_tools.json  @DataDog/ml-observability
tests/snapshots/tests.contrib.anthropic.test_anthropic.test_anthropic_llm_tools_full_use.json  @DataDog/ml-observability
tests/snapshots/tests.contrib.anthropic.test_anthropic.test_anthropic_llm_tools_full_use_stream.json  @DataDog/ml-observability
tests/snapshots/tests.contrib.anthropic.test_anthropic.test_anthropic_llm_tools_stream.json  @DataDog/ml-observability
tests/snapshots/tests.contrib.anthropic.test_anthropic.test_anthropic_llm_tools_stream_helper.json  @DataDog/ml-observability
tests/snapshots/tests.contrib.anthropic.test_anthropic.test_anthropic_llm_unserializable_arg.json  @DataDog/ml-observability

[github-actions]

Bootstrap import analysis

Comparison of import times between this PR and base.

Summary

The average import time from this PR is: 281 ± 3 ms.

The average import time from base is: 286 ± 3 ms.

The import time difference between this PR and base is: -5.1 ± 0.1 ms.

Import time breakdown

The following import paths have shrunk:

ddtrace.auto 2.310 ms (0.82%)

ddtrace.bootstrap.sitecustomize 1.544 ms (0.55%)

ddtrace.bootstrap.preload 1.544 ms (0.55%)

ddtrace.internal.remoteconfig.client 0.714 ms (0.25%)

ddtrace 0.766 ms (0.27%)

ddtrace.internal._unpatched 0.116 ms (0.04%)

subprocess 0.083 ms (0.03%)

contextlib 0.083 ms (0.03%)

json 0.033 ms (0.01%)

json.decoder 0.033 ms (0.01%)

re 0.033 ms (0.01%)

enum 0.033 ms (0.01%)

types 0.033 ms (0.01%)

[pr-commenter]

Benchmarks

Benchmark execution time: 2025-07-01 19:08:50

Comparing candidate commit 7aeb95f in PR branch jsimpher/dac-strip-io-from-anthropic with baseline commit b47c205 in branch main.

Found 0 performance improvements and 0 performance regressions! Performance is the same for 427 metrics, 3 unstable metrics.

[ncybul]

Just a few comments, but glad the integration code was trimmed down so much! Looks great!

[ncybul]

Should this depend on integration.is_pc_sampled_span?

[ncybul]

LGTM