litellm/tests/router_unit_tests/test_router_batch_utils.py
mubashir1osmani 56825926af
fix(vertex/files): stream OpenAI->Vertex batch JSONL uploads (#31036)
* fix(vertex/files): stream OpenAI->Vertex batch JSONL uploads to fix OOM on large files

Large (1GB+) batch JSONL uploads to Vertex AI / GCS caused OOM or killed the worker
because the request body was buffered and multiplied 2-3x in size. The create-file
path is now streaming end-to-end: transform_create_file_request returns a
ResumableChunkedUploadConfig carrying a lazy _OpenAIToVertexBatchUploadStream, and the
HTTP handler opens a GCS resumable session and PUTs the body in bounded 8 MiB chunks
(Content-Range, 308 between chunks) so the transformed payload is never held in full.
The proxy /v1/files endpoint streams from Starlette's spooled upload handle instead of
reading the whole body, and batch rate limiting counts tokens and models in a single
streaming pass.

Only gcs_bucket_name is supported for the GCS target; the legacy bucket_name key is
intentionally not read.

Also removes the unreachable VertexAIFilesHandler create path and everything only it
kept alive (VertexAIJsonlFilesTransformation, _stream_openai_jsonl_to_vertex, the legacy
transform helpers), plus the orphaned batch_utils helpers the streaming rewrite replaced.

* fix(batches): return original JSONL on unparseable row to avoid silent batch truncation

The streaming rewrite of replace_model_in_jsonl accumulated physical lines and
skipped a row on JSONDecodeError to support multi-line objects, but a genuinely
malformed or truncated row never completes: it poisons the buffer, swallows every
following row, and the function still returned the partial rewrite (the rows before
the bad one, already model-rewritten) as if the batch were complete. That turned the
pre-rewrite behavior of returning the original file unchanged (so the provider rejects
the bad batch loudly) into a silent partial submission.

Restore the original-content fallback: when an unparseable remainder is left after the
loop, return the original file_content (rewinding a consumed seekable source) instead of
the truncated output. The multi-line happy path is unchanged.

* test(batches): mock resumable GCS upload in vertex batch prediction test

The vertex batch file-create path now streams to a GCS resumable session via
_aresumable_chunked_upload (httpx send) instead of AsyncHTTPHandler.post, so the
existing test's post mock no longer intercepted the upload and a real request hit
GCS (401). Mock _aresumable_chunked_upload to return the GCS object response; the
resumable protocol itself is covered in test_vertex_ai_files_streaming.py.

* fix(batches): resilient per-row token accounting; no hard-block on count failure

The batch input-file pass iterated a generator whose json.loads raised on a
malformed line; the outer except caught it and stopped the loop, so any body.model
on rows after a bad line was never collected and the model allowlist check ran
against a partial set. It also hard-blocked the batch with a 400 whenever token
counting raised, a backwards-incompatible change from the prior swallow-and-proceed
behavior that breaks legitimate rows the token counter cannot measure (e.g. some
multimodal content).

Iterate the JSONL line-by-line and account each row independently. A malformed line
is skipped (its request cannot run upstream anyway) and a row the counter cannot
measure falls back to a conservative size-based estimate. The loop never aborts, so
the allowlist check always sees every parseable model, and the token total is never
zeroed, so a crafted uncountable row still cannot evade the TPM limit, without
hard-rejecting a legitimate batch.

* perf(vertex/files): unblock async upload; drop empty finalize; widen batch MIME types

Three review follow-ups on the resumable batch upload:
- _aresumable_chunked_upload pulled chunks from a synchronous generator that runs
  the per-row transform inline on the event loop thread, blocking other requests
  between PUTs on large uploads. Each chunk is now produced via asyncio.to_thread.
- _iter_resumable_chunks no longer yields a trailing empty chunk, so an exactly
  chunk-aligned upload finalizes on its last data chunk instead of an extra
  zero-byte PUT; a 0-byte stream still finalizes via the caller's empty request.
- valid_content_type now accepts the MIME types clients label .jsonl batch uploads
  with (text/plain, application/json, ndjson, ...), so such a batch file no longer
  silently bypasses the streaming path into the buffered media upload.

* fix(vertex/files): keep legacy bucket_name as GCS bucket fallback

The rename to gcs_bucket_name dropped the legacy bucket_name key entirely, so an SDK caller passing bucket_name to a Vertex AI file create/retrieve/content call with GCS_BUCKET_NAME unset got ValueError("GCS bucket_name is required") where it previously resolved the bucket. _get_configured_bucket_name now reads gcs_bucket_name, then bucket_name, then the env var, and bucket_name is restored to OPTIONAL_KWARGS_KEYS so it survives get_litellm_params on the retrieve and content paths. gcs_bucket_name keeps precedence when both are present

* style: sort imports in llm_http_handler to satisfy I001 budget

---------

Co-authored-by: Yuneng Jiang <yuneng@berri.ai>
2026-06-24 13:19:57 -07:00

324 lines
12 KiB
Python

import sys
import os
sys.path.insert(
0, os.path.abspath("../..")
) # Adds the parent directory to the system path
import pytest
import json
from io import BytesIO
from typing import Dict, List
from litellm.router_utils.batch_utils import (
replace_model_in_jsonl,
_get_router_metadata_variable_name,
InMemoryFile,
parse_jsonl_with_embedded_newlines,
)
# Fixtures
@pytest.fixture
def sample_jsonl_data() -> List[Dict]:
"""Fixture providing sample JSONL data"""
return [
{
"body": {
"model": "gpt-5-mini",
"messages": [{"role": "user", "content": "Hello"}],
}
},
{"body": {"model": "gpt-5.5", "messages": [{"role": "user", "content": "Hi"}]}},
]
@pytest.fixture
def sample_jsonl_bytes(sample_jsonl_data) -> bytes:
"""Fixture providing sample JSONL as bytes"""
jsonl_str = "\n".join(json.dumps(line) for line in sample_jsonl_data)
return jsonl_str.encode("utf-8")
@pytest.fixture
def sample_file_like(sample_jsonl_bytes):
"""Fixture providing a file-like object"""
return BytesIO(sample_jsonl_bytes)
# Test cases
def test_bytes_input(sample_jsonl_bytes):
"""Test with bytes input"""
new_model = "claude-3"
result = replace_model_in_jsonl(sample_jsonl_bytes, new_model)
assert result is not None
assert isinstance(result, InMemoryFile)
assert result.name == "modified_file.jsonl"
assert result.content_type == "application/jsonl"
def test_tuple_input(sample_jsonl_bytes):
"""Test with tuple input"""
new_model = "claude-3"
test_tuple = ("test.jsonl", sample_jsonl_bytes, "application/json")
result = replace_model_in_jsonl(test_tuple, new_model)
assert result is not None
assert isinstance(result, InMemoryFile)
assert result.name == "modified_file.jsonl"
assert result.content_type == "application/jsonl"
def test_tuple_with_file_handle_rewrites_model(sample_jsonl_bytes):
"""Security regression: when the tuple's content element is a file handle
(batch uploads stream from the spooled upload handle), the model must still
be rewritten. Otherwise a restricted body.model survives unmodified and
bypasses the batch model allowlist, which only checks the upload target."""
new_model = "approved-target-model"
handle = BytesIO(sample_jsonl_bytes)
test_tuple = ("test.jsonl", handle, "application/json")
result = replace_model_in_jsonl(test_tuple, new_model)
assert isinstance(result, InMemoryFile)
rows = [
json.loads(line)
for line in result.getvalue().decode("utf-8").splitlines()
if line.strip()
]
assert rows, "rewrite must produce rows"
# every row now carries the rewritten target, not the original (restricted) model
assert all(row["body"]["model"] == new_model for row in rows)
assert all(row["body"]["model"] != "gpt-5.5" for row in rows)
def test_file_like_object(sample_file_like):
"""Test with file-like object input"""
new_model = "claude-3"
result = replace_model_in_jsonl(sample_file_like, new_model)
assert result is not None
assert isinstance(result, InMemoryFile)
assert result.name == "modified_file.jsonl"
assert result.content_type == "application/jsonl"
def test_router_metadata_variable_name():
"""Test that the variable name is correct"""
assert _get_router_metadata_variable_name(function_name="completion") == "metadata"
assert (
_get_router_metadata_variable_name(function_name="batch") == "litellm_metadata"
)
assert (
_get_router_metadata_variable_name(function_name="acreate_file")
== "litellm_metadata"
)
assert (
_get_router_metadata_variable_name(function_name="aget_file")
== "litellm_metadata"
)
def test_non_json_input():
"""Test that replace_model_in_jsonl returns original content for non-JSON input"""
from litellm.router_utils.batch_utils import replace_model_in_jsonl
# Test with non-JSON string
non_json_str = "This is not a JSON string"
result = replace_model_in_jsonl(non_json_str, "gpt-4")
assert result == non_json_str
# Test with non-JSON bytes
non_json_bytes = b"This is not JSON bytes"
result = replace_model_in_jsonl(non_json_bytes, "gpt-4")
assert result == non_json_bytes
# Test with non-JSON file-like object
from io import BytesIO
non_json_file = BytesIO(b"This is not JSON in a file")
result = replace_model_in_jsonl(non_json_file, "gpt-4")
assert result == non_json_file
def test_should_replace_model_in_jsonl():
"""Test that should_replace_model_in_jsonl returns the correct value"""
from litellm.router_utils.batch_utils import should_replace_model_in_jsonl
assert should_replace_model_in_jsonl(purpose="batch") is True
assert should_replace_model_in_jsonl(purpose="test") is False
assert should_replace_model_in_jsonl(purpose="user_data") is False
def test_parse_jsonl_with_embedded_newlines_simple():
"""Test parsing simple JSONL without embedded newlines"""
content = '{"id": 1, "name": "test"}\n{"id": 2, "name": "test2"}'
result = parse_jsonl_with_embedded_newlines(content)
assert len(result) == 2
assert result[0] == {"id": 1, "name": "test"}
assert result[1] == {"id": 2, "name": "test2"}
def test_parse_jsonl_with_embedded_newlines_in_strings():
"""Test parsing JSONL with newlines embedded in string values"""
content = (
'{"id": 1, "message": "Line 1\\nLine 2\\nLine 3"}\n{"id": 2, "message": "test"}'
)
result = parse_jsonl_with_embedded_newlines(content)
assert len(result) == 2
assert result[0] == {"id": 1, "message": "Line 1\nLine 2\nLine 3"}
assert result[1] == {"id": 2, "message": "test"}
def test_parse_jsonl_with_embedded_newlines_real_world_example():
"""Test with the real-world example from the Cooler Master Shark X case"""
# This simulates the actual problem case from the user's log
content = """{"custom_id":"16546277850245725","method":"POST","url":"/v1/chat/completions","body":{"model":"openai-gpt-4o-mini-dp-items-translation-dag","messages":[{"role":"system","content":"Translate the product title and description for an e-commerce marketplace in Saudi Arabia and the UAE. Text may be in English or Arabic.\\n"},{"role":"user","content":"\\nOriginal Title: ```Cooler Master Shark X PC Case```\\nOriginal Description: ```UNIQUE MASTERPIECEShark X is a system that provides an impressive unique alternative to traditional PC systems. Shark X will stand out and can be the ultimate trophy or conversation piece for people looking for a unique setup that stands head and fins above the res.```\\nStore Name: ```geekay```\\n"}]}}"""
result = parse_jsonl_with_embedded_newlines(content)
assert len(result) == 1
assert result[0]["custom_id"] == "16546277850245725"
assert result[0]["method"] == "POST"
assert result[0]["body"]["model"] == "openai-gpt-4o-mini-dp-items-translation-dag"
assert len(result[0]["body"]["messages"]) == 2
assert "Translate the product title" in result[0]["body"]["messages"][0]["content"]
assert (
"Cooler Master Shark X PC Case" in result[0]["body"]["messages"][1]["content"]
)
assert "UNIQUE MASTERPIECEShark X" in result[0]["body"]["messages"][1]["content"]
def test_parse_jsonl_with_embedded_newlines_multiple_complex_objects():
"""Test parsing multiple complex JSON objects with embedded newlines"""
content = """{"id":1,"text":"Line 1\\nLine 2"}
{"id":2,"nested":{"field":"Value\\nWith\\nNewlines"}}
{"id":3,"simple":"test"}"""
result = parse_jsonl_with_embedded_newlines(content)
assert len(result) == 3
assert result[0]["id"] == 1
assert result[0]["text"] == "Line 1\nLine 2"
assert result[1]["id"] == 2
assert result[1]["nested"]["field"] == "Value\nWith\nNewlines"
assert result[2]["id"] == 3
assert result[2]["simple"] == "test"
def test_parse_jsonl_with_embedded_newlines_no_trailing_newline():
"""Test parsing JSONL without trailing newline"""
content = '{"id": 1, "name": "test"}'
result = parse_jsonl_with_embedded_newlines(content)
assert len(result) == 1
assert result[0] == {"id": 1, "name": "test"}
def test_parse_jsonl_with_embedded_newlines_empty_string():
"""Test parsing empty string"""
content = ""
result = parse_jsonl_with_embedded_newlines(content)
assert len(result) == 0
def test_parse_jsonl_with_embedded_newlines_whitespace_only():
"""Test parsing whitespace-only content"""
content = " \n \n "
result = parse_jsonl_with_embedded_newlines(content)
assert len(result) == 0
def test_replace_model_in_jsonl_malformed_middle_row_returns_original():
"""Regression: a malformed/truncated middle row must not silently drop the
rows that follow it. The streaming rewrite accumulates physical lines into a
buffer; a row that never parses poisons the buffer so every later valid row
is concatenated into it and dropped. Returning that partial rewrite would
ship a truncated batch with no error to the caller. Instead the original
content is returned unchanged so the provider rejects the bad batch loudly."""
content = (
b'{"custom_id":"a","body":{"model":"x"}}\n'
b'{"custom_id":"b","body":{"model":\n' # truncated, never completes
b'{"custom_id":"c","body":{"model":"x"}}\n'
)
result = replace_model_in_jsonl(content, "new-model")
assert (
result == content
), "must return the original unchanged, not a partial rewrite"
def test_replace_model_in_jsonl_malformed_row_seekable_handle_rewound():
"""When the source is a seekable handle that gets consumed during the failed
rewrite, it must be rewound to 0 so the caller can re-read the full original."""
content = (
b'{"custom_id":"a","body":{"model":"x"}}\n'
b'{"custom_id":"b","body":{"model":\n'
b'{"custom_id":"c","body":{"model":"x"}}\n'
)
handle = BytesIO(content)
result = replace_model_in_jsonl(handle, "new-model")
assert result is handle
assert handle.read() == content, "handle must be rewound for the caller to re-read"
def test_replace_model_in_jsonl_multi_row_rewrites_every_model():
"""Happy path: a well-formed multi-row file gets every row's model rewritten
and no row is dropped."""
content = (
b'{"custom_id":"a","body":{"model":"old1"}}\n'
b'{"custom_id":"b","body":{"model":"old2"}}\n'
b'{"custom_id":"c","body":{"model":"old3"}}\n'
)
result = replace_model_in_jsonl(content, "new-model")
assert isinstance(result, InMemoryFile)
rows = [
json.loads(line)
for line in result.getvalue().decode("utf-8").splitlines()
if line.strip()
]
assert [row["custom_id"] for row in rows] == ["a", "b", "c"]
assert all(row["body"]["model"] == "new-model" for row in rows)
def test_replace_model_in_jsonl_with_embedded_newlines():
"""Test that replace_model_in_jsonl works correctly with embedded newlines in content"""
# Create a JSONL with embedded newlines in the message content
jsonl_data = {
"custom_id": "test123",
"body": {
"model": "old-model",
"messages": [
{"role": "user", "content": "This is a message\nwith multiple\nlines"}
],
},
}
jsonl_bytes = json.dumps(jsonl_data).encode("utf-8")
new_model = "new-model"
result = replace_model_in_jsonl(jsonl_bytes, new_model)
assert isinstance(result, InMemoryFile)
# Read and parse the result
result_content = result.read().decode("utf-8")
result_json = json.loads(result_content)
# Verify the model was replaced
assert result_json["body"]["model"] == "new-model"
# Verify the content with newlines is preserved
assert (
result_json["body"]["messages"][0]["content"]
== "This is a message\nwith multiple\nlines"
)
assert result_json["custom_id"] == "test123"