admin/openhuman
1
0
Fork 0
mirror of https://github.com/tinyhumansai/openhuman.git synced 2026-07-27 21:08:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix(inference): cap extraction max_tokens + classify provider 402 to stop the retry/Sentry storm (#3616) (#3617)

Browse Source 
Co-authored-by: Steven Enamakel <enamakel@tinyhumans.ai>
This commit is contained in:
oxoxDev
2026-06-15 18:27:39 -07:00
committed by GitHub
co-authored by Steven Enamakel
parent 5b80f25cd4
commit 445b11afc2
35 changed files with 707 additions and 43 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/inference_agent_raw_coverage_e2e.rs
+6
View File
@@ -2183,6 +2183,7 @@ async fn inference_provider_trait_defaults_cover_prompt_guided_paths() {
messages: &[ChatMessage::user("need docs")],
tools: Some(&[tool_spec.clone()]),
stream: None,
max_tokens: None,
},
"agentic-v1",
0.4,
@@ -2198,6 +2199,7 @@ async fn inference_provider_trait_defaults_cover_prompt_guided_paths() {
messages: &[ChatMessage::user("plain")],
tools: None,
stream: None,
max_tokens: None,
},
"agentic-v1",
0.5,
@@ -2283,6 +2285,7 @@ async fn inference_openai_compatible_provider_covers_native_streaming_and_fallba
],
tools: Some(&[tool_spec.clone(), tool_spec.clone()]),
stream: Some(&delta_tx),
max_tokens: None,
},
"stream-native",
0.9,
@@ -2322,6 +2325,7 @@ async fn inference_openai_compatible_provider_covers_native_streaming_and_fallba
messages: &[ChatMessage::user("json encoded tool call")],
tools: None,
stream: None,
max_tokens: None,
},
"tool-content-json",
0.2,
@@ -4429,6 +4433,7 @@ async fn inference_router_provider_covers_hint_tier_and_passthrough_routing() {
messages: &[ChatMessage::user("fallback")],
tools: None,
stream: None,
max_tokens: None,
},
"reasoning-v1",
0.4,
@@ -4548,6 +4553,7 @@ async fn inference_reliable_provider_covers_retry_fallback_and_aggregate_errors(
messages: &[ChatMessage::user("fail")],
tools: None,
stream: None,
max_tokens: None,
},
"missing-model",
0.0,
tests/inference_compatible_admin_leftovers_raw_coverage_e2e.rs
+3
View File
@@ -141,6 +141,7 @@ async fn compatible_native_leftovers_cover_tool_history_function_call_and_stream
],
tools: Some(&tools),
stream: None,
max_tokens: None,
},
"function-call-model",
0.4,
@@ -165,6 +166,7 @@ async fn compatible_native_leftovers_cover_tool_history_function_call_and_stream
messages: &[ChatMessage::user("content json")],
tools: Some(&tools),
stream: None,
max_tokens: None,
},
"content-json-tools",
0.4,
@@ -186,6 +188,7 @@ async fn compatible_native_leftovers_cover_tool_history_function_call_and_stream
messages: &[ChatMessage::user("stream ordering")],
tools: Some(&tools),
stream: Some(&tx),
max_tokens: None,
},
"stream-out-of-order-tool",
0.4,
tests/inference_compatible_matrix_raw_coverage_e2e.rs
+6
View File
@@ -154,6 +154,7 @@ async fn openai_compatible_matrix_covers_auth_requests_responses_and_streaming()
],
tools: Some(&tools),
stream: None,
max_tokens: None,
},
"native-tools",
0.2,
@@ -181,6 +182,7 @@ async fn openai_compatible_matrix_covers_auth_requests_responses_and_streaming()
messages: &[ChatMessage::user("stream")],
tools: Some(&tools),
stream: Some(&tx),
max_tokens: None,
},
"stream-sse",
0.3,
@@ -212,6 +214,7 @@ async fn openai_compatible_matrix_covers_auth_requests_responses_and_streaming()
messages: &[ChatMessage::user("json stream")],
tools: None,
stream: Some(&json_tx),
max_tokens: None,
},
"stream-json",
0.3,
@@ -229,6 +232,7 @@ async fn openai_compatible_matrix_covers_auth_requests_responses_and_streaming()
messages: &[ChatMessage::user("retry without tools")],
tools: Some(&tools),
stream: Some(&retry_tx),
max_tokens: None,
},
"stream-tools-unsupported",
0.3,
@@ -350,6 +354,7 @@ async fn compatible_error_matrix_covers_status_malformed_and_no_fallback_paths()
messages: &[ChatMessage::user("stream fail")],
tools: None,
stream: Some(&sse_tx),
max_tokens: None,
},
"stream-status-error",
0.1,
@@ -407,6 +412,7 @@ async fn ollama_compatible_matrix_covers_authless_chat_and_streaming_errors() {
messages: &[ChatMessage::user("ollama stream")],
tools: None,
stream: Some(&tx),
max_tokens: None,
},
"ollama-stream",
0.0,
tests/inference_local_admin_raw_coverage_e2e.rs
+1
View File
@@ -151,6 +151,7 @@ async fn compatible_provider_covers_retry_headers_responses_and_parse_errors() {
],
tools: Some(&tools),
stream: Some(&tx),
max_tokens: None,
},
"stream-tools-unsupported",
0.2,
tests/inference_provider_admin_round22_raw_coverage_e2e.rs
+1
View File
@@ -412,6 +412,7 @@ async fn reliable_provider_covers_chat_tools_streaming_and_context_bail_edges()
messages: &[ChatMessage::user("retry me")],
tools: None,
stream: None,
max_tokens: None,
},
"retry-model",
0.2,
tests/inference_provider_raw_coverage_e2e.rs
+2
View File
@@ -112,6 +112,7 @@ async fn compatible_provider_covers_chat_responses_streaming_tools_and_errors()
messages: &[ChatMessage::user("use a tool")],
tools: Some(&tools),
stream: None,
max_tokens: None,
},
"tool-model",
0.1,
@@ -135,6 +136,7 @@ async fn compatible_provider_covers_chat_responses_streaming_tools_and_errors()
messages: &[ChatMessage::user("stream it")],
tools: Some(&tools),
stream: Some(&tx),
max_tokens: None,
},
"stream-model",
0.1,
tests/inference_round26_raw_coverage_e2e.rs
+3
View File
@@ -111,6 +111,7 @@ async fn compatible_streaming_covers_tool_deltas_json_fallback_and_retry_without
messages: &messages,
tools: Some(&tools),
stream: Some(&delta_tx),
max_tokens: None,
},
"stream-tools",
0.4,
@@ -154,6 +155,7 @@ async fn compatible_streaming_covers_tool_deltas_json_fallback_and_retry_without
messages: &[ChatMessage::user("json stream fallback")],
tools: None,
stream: Some(&json_tx),
max_tokens: None,
},
"json-stream",
0.2,
@@ -170,6 +172,7 @@ async fn compatible_streaming_covers_tool_deltas_json_fallback_and_retry_without
messages: &[ChatMessage::user("retry without tools")],
tools: Some(&[tool_spec("lookup")]),
stream: Some(&retry_tx),
max_tokens: None,
},
"tool-retry",
0.2,
tests/owned_domain_raw_coverage_e2e.rs
+3
View File
@@ -362,6 +362,7 @@ async fn openai_compatible_provider_covers_auth_temperature_tool_fallback_and_re
messages: &messages,
tools: Some(&[tool.clone(), tool]),
stream: None,
max_tokens: None,
},
"gpt-5-mini",
0.6,
@@ -383,6 +384,7 @@ async fn openai_compatible_provider_covers_auth_temperature_tool_fallback_and_re
messages: &messages,
tools: None,
stream: None,
max_tokens: None,
},
"gpt-5-mini",
0.6,
@@ -450,6 +452,7 @@ async fn openai_compatible_provider_streaming_json_fallback_aggregates_response(
messages: &messages,
tools: None,
stream: Some(&tx),
max_tokens: None,
},
"stream-model",
0.7,