fix(openai): allow token inputs for Embeddings endpoint [backport #5890 to 1.13] (#5909)

Backports #5890 to 1.13.

Fixes #5884.

This PR adds handling for non-string request input params for the
`Embeddings.create()` method. This API method accepts strings, array of
strings, as well as token arrays (list of ints) and arrays of token
arrays (list of list of ints), but the previous implementation only
accounted for strings and array of strings. Token arrays or arrays of
token arrays being passed in as arguments would then cause errors as the
traced request handler for the Embeddings endpoint did not account for
non-string types.

## Checklist

- [x] Change(s) are motivated and described in the PR description.
- [x] Testing strategy is described if automated tests are not included
in the PR.
- [x] Risk is outlined (performance impact, potential for breakage,
maintainability, etc).
- [x] Change is maintainable (easy to change, telemetry, documentation).
- [x] [Library release note
guidelines](https://ddtrace.readthedocs.io/en/stable/contributing.html#Release-Note-Guidelines)
are followed.
- [x] Documentation is included (in-code, generated user docs, [public
corp docs](https://github.com/DataDog/documentation/)).
- [x] OPTIONAL: PR description includes explicit acknowledgement of the
performance implications of the change as reported in the benchmarks PR
comment.

## Reviewer Checklist

- [x] Title is accurate.
- [x] No unnecessary changes are introduced.
- [x] Description motivates each change.
- [x] Avoids breaking
[API](https://ddtrace.readthedocs.io/en/stable/versioning.html#interfaces)
changes unless absolutely necessary.
- [x] Testing strategy adequately addresses listed risk(s).
- [x] Change is maintainable (easy to change, telemetry, documentation).
- [x] Release note makes sense to a user of the library.
- [x] Reviewer has explicitly acknowledged and discussed the performance
implications of this PR as reported in the benchmarks PR comment.

Author: Yun-Kim

ddtrace/contrib/openai/patch.py

@@ -630,7 +630,7 @@ def handle_request(self, pin, integration, span, args, kwargs):
                 if kw_attr == "input" and integration.is_pc_sampled_span(span):
                     if isinstance(kwargs["input"], list):
                         for idx, inp in enumerate(kwargs["input"]):
-                            span.set_tag_str("openai.request.input.%d" % idx, integration.trunc(inp))
+                            span.set_tag_str("openai.request.input.%d" % idx, integration.trunc(str(inp)))
                     else:
                         span.set_tag("openai.request.%s" % kw_attr, kwargs[kw_attr])
                 else:

docs/spelling_wordlist.txt

@@ -133,6 +133,7 @@ mysqlclient
 mysqldb
 namespace
 obfuscator
+openai
 opensearch
 opentracer
 opentracing

releasenotes/notes/fix-openai-embeddings-input-7feb1198fabac31b.yaml

@@ -0,0 +1,5 @@
+---
+fixes:
+  - |
+    openai: Resolves an issue where using an array of tokens or an array of token arrays 
+    for the Embeddings endpoint caused an AttributeError.

tests/contrib/openai/test_openai.py

@@ -546,6 +546,32 @@ def test_embedding(api_key_in_env, request_api_key, openai, openai_vcr, snapshot
         openai.Embedding.create(api_key=request_api_key, input="hello world", model="text-embedding-ada-002")
 
 
+@pytest.mark.snapshot(ignores=["meta.http.useragent"])
+def test_embedding_string_array(openai, openai_vcr, snapshot_tracer):
+    if not hasattr(openai, "Embedding"):
+        pytest.skip("embedding not supported for this version of openai")
+    with openai_vcr.use_cassette("embedding.yaml"):
+        openai.Embedding.create(input=["hello world", "hello again"], model="text-embedding-ada-002")
+
+
+@pytest.mark.snapshot(ignores=["meta.http.useragent"])
+def test_embedding_token_array(openai, openai_vcr, snapshot_tracer):
+    if not hasattr(openai, "Embedding"):
+        pytest.skip("embedding not supported for this version of openai")
+    with openai_vcr.use_cassette("embedding.yaml"):
+        openai.Embedding.create(input=[1111, 2222, 3333], model="text-embedding-ada-002")
+
+
+@pytest.mark.snapshot(ignores=["meta.http.useragent"])
+def test_embedding_array_of_token_arrays(openai, openai_vcr, snapshot_tracer):
+    if not hasattr(openai, "Embedding"):
+        pytest.skip("embedding not supported for this version of openai")
+    with openai_vcr.use_cassette("embedding.yaml"):
+        openai.Embedding.create(
+            input=[[1111, 2222, 3333], [4444, 5555, 6666], [7777, 8888, 9999]], model="text-embedding-ada-002"
+        )
+
+
 @pytest.mark.asyncio
 @pytest.mark.snapshot(ignores=["meta.http.useragent"])
 @pytest.mark.parametrize("api_key_in_env", [True, False])

tests/snapshots/tests.contrib.openai.test_openai.test_embedding_array_of_token_arrays.json

@@ -0,0 +1,41 @@
+[[
+  {
+    "name": "openai.request",
+    "service": "",
+    "resource": "embeddings/text-embedding-ada-002",
+    "trace_id": 0,
+    "span_id": 1,
+    "parent_id": 0,
+    "type": "",
+    "error": 0,
+    "meta": {
+      "_dd.p.dm": "-0",
+      "api_base": "https://api.openai.com/v1",
+      "component": "openai",
+      "language": "python",
+      "openai.endpoint": "embeddings",
+      "openai.model": "text-embedding-ada-002",
+      "openai.organization.name": "datadog-4",
+      "openai.organization.ratelimit.requests.remaining": "2999",
+      "openai.request.input.0": "[1111, 2222, 3333]",
+      "openai.request.input.1": "[4444, 5555, 6666]",
+      "openai.request.input.2": "[7777, 8888, 9999]",
+      "openai.request.model": "text-embedding-ada-002",
+      "openai.user.api_key": "sk-...key>",
+      "runtime-id": "b168eb19ef14414ca786ac99826ef9e0"
+    },
+    "metrics": {
+      "_dd.agent_psr": 1.0,
+      "_dd.measured": 1,
+      "_dd.top_level": 1,
+      "_dd.tracer_kr": 1.0,
+      "_sampling_priority_v1": 1,
+      "openai.response.data.embedding-length": 1536,
+      "openai.response.data.num-embeddings": 1,
+      "openai.response.usage.prompt_tokens": 2,
+      "openai.response.usage.total_tokens": 2,
+      "process_id": 85925
+    },
+    "duration": 3655000,
+    "start": 1684349256345286000
+  }]]

tests/snapshots/tests.contrib.openai.test_openai.test_embedding_string_array.json

@@ -0,0 +1,40 @@
+[[
+  {
+    "name": "openai.request",
+    "service": "",
+    "resource": "embeddings/text-embedding-ada-002",
+    "trace_id": 0,
+    "span_id": 1,
+    "parent_id": 0,
+    "type": "",
+    "error": 0,
+    "meta": {
+      "_dd.p.dm": "-0",
+      "api_base": "https://api.openai.com/v1",
+      "component": "openai",
+      "language": "python",
+      "openai.endpoint": "embeddings",
+      "openai.model": "text-embedding-ada-002",
+      "openai.organization.name": "datadog-4",
+      "openai.organization.ratelimit.requests.remaining": "2999",
+      "openai.request.input.0": "hello world",
+      "openai.request.input.1": "hello again",
+      "openai.request.model": "text-embedding-ada-002",
+      "openai.user.api_key": "sk-...key>",
+      "runtime-id": "b168eb19ef14414ca786ac99826ef9e0"
+    },
+    "metrics": {
+      "_dd.agent_psr": 1.0,
+      "_dd.measured": 1,
+      "_dd.top_level": 1,
+      "_dd.tracer_kr": 1.0,
+      "_sampling_priority_v1": 1,
+      "openai.response.data.embedding-length": 1536,
+      "openai.response.data.num-embeddings": 1,
+      "openai.response.usage.prompt_tokens": 2,
+      "openai.response.usage.total_tokens": 2,
+      "process_id": 85925
+    },
+    "duration": 3517000,
+    "start": 1684349256264173000
+  }]]

tests/snapshots/tests.contrib.openai.test_openai.test_embedding_token_array.json

@@ -0,0 +1,41 @@
+[[
+  {
+    "name": "openai.request",
+    "service": "",
+    "resource": "embeddings/text-embedding-ada-002",
+    "trace_id": 0,
+    "span_id": 1,
+    "parent_id": 0,
+    "type": "",
+    "error": 0,
+    "meta": {
+      "_dd.p.dm": "-0",
+      "api_base": "https://api.openai.com/v1",
+      "component": "openai",
+      "language": "python",
+      "openai.endpoint": "embeddings",
+      "openai.model": "text-embedding-ada-002",
+      "openai.organization.name": "datadog-4",
+      "openai.organization.ratelimit.requests.remaining": "2999",
+      "openai.request.input.0": "1111",
+      "openai.request.input.1": "2222",
+      "openai.request.input.2": "3333",
+      "openai.request.model": "text-embedding-ada-002",
+      "openai.user.api_key": "sk-...key>",
+      "runtime-id": "b168eb19ef14414ca786ac99826ef9e0"
+    },
+    "metrics": {
+      "_dd.agent_psr": 1.0,
+      "_dd.measured": 1,
+      "_dd.top_level": 1,
+      "_dd.tracer_kr": 1.0,
+      "_sampling_priority_v1": 1,
+      "openai.response.data.embedding-length": 1536,
+      "openai.response.data.num-embeddings": 1,
+      "openai.response.usage.prompt_tokens": 2,
+      "openai.response.usage.total_tokens": 2,
+      "process_id": 85925
+    },
+    "duration": 3522000,
+    "start": 1684349256312922000
+  }]]