Build AI methods with your coding agent, then run them anywhere: as an MCP for chatbots, as a webapp for people, or via API for your software.
Release v0.83.0
Louis Choquel08e2f9edab19
diff --git a/.badges/tests.json b/.badges/tests.jsonindex 8a5d7f402..5f95578fa 100644--- a/.badges/tests.json+++ b/.badges/tests.json@@ -1,7 +1,7 @@{"schemaVersion": 1,"label": "tests",- "message": "17906",+ "message": "18033","color": "blue","cacheSeconds": 300}diff --git a/.test_durations b/.test_durationsindex 09e88df22..ef073cbb2 100644--- a/.test_durations+++ b/.test_durations@@ -282,6 +282,7 @@"tests/integration/pipelex/cli/test_run_bundle_verdict_cli.py::TestRunBundleVerdictCli::test_run_bundle_renders_the_invalid_bundle_panel[directory]": 0.1654,"tests/integration/pipelex/cli/test_run_bundle_verdict_cli.py::TestRunBundleVerdictCli::test_run_bundle_renders_the_invalid_bundle_panel[file]": 0.0074,"tests/integration/pipelex/cli/test_run_located_failure_panel.py::TestRunLocatedFailurePanel::test_unserved_model_renders_the_model_panel_at_the_failing_step": 0.1598,+ "tests/integration/pipelex/cli/test_run_truncated_completion.py::TestRunTruncatedCompletion::test_a_truncated_text_fails_the_run_naming_the_pipe_and_the_error": 0.9207,"tests/integration/pipelex/cli/test_save_csv_empty_result_row_model.py::TestSaveCsvEmptyResultRowModel::test_the_reload_resolves_a_generated_structure_class": 0.48,"tests/integration/pipelex/cli/test_trace_input_semantics_cmd.py::TestTraceInputSemantics::test_per_hop_captures_on_small_bundle": 1.5909,"tests/integration/pipelex/cli/test_trace_input_semantics_cmd.py::TestTraceInputSemantics::test_probe_fixture_validates_and_traces": 0.6866,@@ -5508,6 +5509,57 @@"tests/unit/pipelex/cogt/judgment/test_judgment_worker_file_capability.py::TestJudgmentWorkerFileCapability::test_a_text_only_model_refuses_an_image": 0.0927,"tests/unit/pipelex/cogt/judgment/test_judgment_worker_file_capability.py::TestJudgmentWorkerFileCapability::test_an_image_input_carrying_a_document_is_refused": 0.001,"tests/unit/pipelex/cogt/judgment/test_judgment_worker_file_capability.py::TestJudgmentWorkerFileCapability::test_more_images_than_the_model_reads_are_refused": 0.0012,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_a_gemini_value_is_read_whatever_its_case": 0.0003,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_a_missing_value_is_normal_without_a_word": 0.0005,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_a_normal_stop_raises_nothing": 0.0005,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_a_refused_stop_raises_the_refusal": 0.0006,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_a_responses_answer_is_classified[completed-completed-None-normal]": 0.0004,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_a_responses_answer_is_classified[failed-failed-None-normal]": 0.0004,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_a_responses_answer_is_classified[filtered-incomplete-content_filter-refused]": 0.0005,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_a_responses_answer_is_classified[incomplete with no reason-incomplete-None-truncated]": 0.0017,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_a_responses_answer_is_classified[no status-None-None-normal]": 0.0007,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_a_responses_answer_is_classified[truncated-incomplete-max_output_tokens-truncated]": 0.1172,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_a_truncated_stop_raises_the_truncation_naming_its_limit[context_window_exceeded-model_context_window_exceeded-context_window]": 0.0004,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_a_truncated_stop_raises_the_truncation_naming_its_limit[truncated-max_tokens-max_tokens]": 0.0009,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_an_incomplete_answer_without_a_known_reason_is_logged_and_taken_as_truncated[None]": 0.0006,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_an_incomplete_answer_without_a_known_reason_is_logged_and_taken_as_truncated[brand_new_reason]": 0.0007,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_an_unknown_gemini_block_reason_is_logged_and_taken_as_a_refusal": 0.0011,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_an_unknown_value_is_logged_and_taken_as_normal": 0.0006,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_covers_its_sdk_type": 0.1115,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[anthropic context window-model_context_window_exceeded-context_window_exceeded]": 0.0004,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[anthropic normal-end_turn-normal]": 0.0006,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[anthropic paused-pause_turn-normal]": 0.1194,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[anthropic refused-refusal-refused]": 0.0004,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[anthropic truncated-max_tokens-truncated]": 0.0004,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[converse context window-model_context_window_exceeded-context_window_exceeded]": 0.0006,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[converse filtered-content_filtered-refused]": 0.0009,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[converse guardrail-guardrail_intervened-refused]": 0.1172,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[converse normal-stop_sequence-normal]": 0.0005,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[converse truncated-max_tokens-truncated]": 0.0005,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[gateway strict refusal-refusal-refused]": 0.0004,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[gemini blocklist-BLOCKLIST-refused]": 0.0004,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[gemini normal-STOP-normal]": 0.0004,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[gemini privacy-SPII-refused]": 0.0007,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[gemini prohibited-PROHIBITED_CONTENT-refused]": 0.0004,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[gemini recitation-RECITATION-refused]": 0.0005,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[gemini safety-SAFETY-refused]": 0.0007,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[gemini truncated-MAX_TOKENS-truncated]": 0.0005,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[mistral context window-model_length-context_window_exceeded]": 0.0004,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[mistral normal-stop-normal]": 0.1153,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[mistral truncated-length-truncated]": 0.0007,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[openai filtered-content_filter-refused]": 0.0951,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[openai normal-stop-normal]": 0.0852,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[openai tool call-tool_calls-normal]": 0.0975,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_each_vocabulary_is_classified[openai truncated-length-truncated]": 0.0006,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_every_gemini_block_reason_is_a_refusal[BLOCKED_REASON_UNSPECIFIED]": 0.0005,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_every_gemini_block_reason_is_a_refusal[BLOCKLIST]": 0.0006,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_every_gemini_block_reason_is_a_refusal[IMAGE_SAFETY]": 0.0005,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_every_gemini_block_reason_is_a_refusal[JAILBREAK]": 0.0005,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_every_gemini_block_reason_is_a_refusal[MODEL_ARMOR]": 0.0004,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_every_gemini_block_reason_is_a_refusal[OTHER]": 0.1233,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_every_gemini_block_reason_is_a_refusal[PROHIBITED_CONTENT]": 0.0006,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_every_gemini_block_reason_is_a_refusal[SAFETY]": 0.0006,+ "tests/unit/pipelex/cogt/llm/test_completion_stop.py::TestCompletionStop::test_the_vocabularies_agree_on_every_shared_value": 0.0004,"tests/unit/pipelex/cogt/llm/test_instructor_retry.py::TestInstructorSchemaRetrying::test_json_decode_error_is_retried": 0.0047,"tests/unit/pipelex/cogt/llm/test_instructor_retry.py::TestInstructorSchemaRetrying::test_predicate_matches_what_instructor_itself_reasks": 0.0023,"tests/unit/pipelex/cogt/llm/test_instructor_retry.py::TestInstructorSchemaRetrying::test_transport_error_is_not_retried[transport_exc0]": 0.0022,@@ -5636,6 +5688,10 @@"tests/unit/pipelex/cogt/llm/test_llm_worker_request_check.py::TestLLMWorkerRequestCheck::test_the_shared_rule_passes_every_other_request[adaptive_budget]": 0.0006,"tests/unit/pipelex/cogt/llm/test_llm_worker_request_check.py::TestLLMWorkerRequestCheck::test_the_shared_rule_passes_every_other_request[manual_effort]": 0.0323,"tests/unit/pipelex/cogt/llm/test_llm_worker_request_check.py::TestLLMWorkerRequestCheck::test_the_shared_rule_passes_every_other_request[no_thinking_no_setting]": 0.0173,+ "tests/unit/pipelex/cogt/llm/test_llm_worker_stopped_text_usage.py::TestLLMWorkerStoppedTextUsage::test_a_failure_with_no_usage_recorded_reports_nothing": 0.8856,+ "tests/unit/pipelex/cogt/llm/test_llm_worker_stopped_text_usage.py::TestLLMWorkerStoppedTextUsage::test_a_returned_text_is_reported_once": 0.1283,+ "tests/unit/pipelex/cogt/llm/test_llm_worker_stopped_text_usage.py::TestLLMWorkerStoppedTextUsage::test_a_stopped_text_raises_and_reports_its_billed_usage[content_filter-LLMCompletionRefusedError]": 0.0164,+ "tests/unit/pipelex/cogt/llm/test_llm_worker_stopped_text_usage.py::TestLLMWorkerStoppedTextUsage::test_a_stopped_text_raises_and_reports_its_billed_usage[length-LLMCompletionTruncatedError]": 0.7599,"tests/unit/pipelex/cogt/llm/test_structured_output.py::TestStructureMethod::test_as_instructor_mode_is_the_core_mode_instructor_resolves_it_to[instructor/anthropic_json]": 0.0004,"tests/unit/pipelex/cogt/llm/test_structured_output.py::TestStructureMethod::test_as_instructor_mode_is_the_core_mode_instructor_resolves_it_to[instructor/anthropic_reasoning_tools]": 0.0004,"tests/unit/pipelex/cogt/llm/test_structured_output.py::TestStructureMethod::test_as_instructor_mode_is_the_core_mode_instructor_resolves_it_to[instructor/anthropic_tools]": 0.0011,@@ -6742,6 +6798,18 @@"tests/unit/pipelex/cogt/test_inference_error_domain.py::TestInferenceErrorDomain::test_uncategorized_cogt_error_still_has_no_domain": 0.0017,"tests/unit/pipelex/cogt/test_inference_error_domain.py::TestInferenceErrorDomain::test_uncategorized_cogt_error_still_inherits_the_cause_domain": 0.0004,"tests/unit/pipelex/cogt/test_inference_error_domain.py::TestInferenceErrorDomain::test_unknown_asserts_no_domain": 0.0004,+ "tests/unit/pipelex/cogt/test_llm_completion_stop_errors.py::TestLLMCompletionStopErrors::test_a_context_window_stop_advises_a_shorter_input_never_a_higher_max_tokens": 0.0006,+ "tests/unit/pipelex/cogt/test_llm_completion_stop_errors.py::TestLLMCompletionStopErrors::test_a_cut_with_no_max_tokens_sent_advises_setting_one_rather_than_raising_it": 0.1013,+ "tests/unit/pipelex/cogt/test_llm_completion_stop_errors.py::TestLLMCompletionStopErrors::test_a_run_failure_located_at_the_pipe_reports_the_truncation": 0.0006,+ "tests/unit/pipelex/cogt/test_llm_completion_stop_errors.py::TestLLMCompletionStopErrors::test_the_message_survives_strict_disclosure_without_the_partial_text[error0]": 0.0004,+ "tests/unit/pipelex/cogt/test_llm_completion_stop_errors.py::TestLLMCompletionStopErrors::test_the_message_survives_strict_disclosure_without_the_partial_text[error1]": 0.0004,+ "tests/unit/pipelex/cogt/test_llm_completion_stop_errors.py::TestLLMCompletionStopErrors::test_the_message_survives_strict_disclosure_without_the_partial_text[error2]": 0.0004,+ "tests/unit/pipelex/cogt/test_llm_completion_stop_errors.py::TestLLMCompletionStopErrors::test_the_refusal_names_the_pipe_model_stop_and_next_step": 0.0978,+ "tests/unit/pipelex/cogt/test_llm_completion_stop_errors.py::TestLLMCompletionStopErrors::test_the_report_carries_the_fields_an_agent_branches_on[LLMCompletionRefusedError-error2-Revise the pipe's prompt or its input.]": 0.1202,+ "tests/unit/pipelex/cogt/test_llm_completion_stop_errors.py::TestLLMCompletionStopErrors::test_the_report_carries_the_fields_an_agent_branches_on[LLMCompletionTruncatedError-error0-Raise the pipe's max_tokens, or shorten its input.]": 0.0007,+ "tests/unit/pipelex/cogt/test_llm_completion_stop_errors.py::TestLLMCompletionStopErrors::test_the_report_carries_the_fields_an_agent_branches_on[LLMCompletionTruncatedError-error1-Shorten the pipe's input, lower its reasoning effort, or choose a model with a larger context window.]": 0.0006,+ "tests/unit/pipelex/cogt/test_llm_completion_stop_errors.py::TestLLMCompletionStopErrors::test_the_truncation_names_the_pipe_model_stop_counts_and_next_step": 0.0882,+ "tests/unit/pipelex/cogt/test_llm_completion_stop_errors.py::TestLLMCompletionStopErrors::test_without_a_pipe_or_counts_the_messages_still_read_as_sentences": 0.0009,"tests/unit/pipelex/cogt/test_model_choice_not_found_disclosure.py::TestModelChoiceNotFoundDisclosure::test_message_survives_strict_disclosure": 0.771,"tests/unit/pipelex/cogt/test_model_choice_not_found_disclosure.py::TestModelChoiceNotFoundDisclosure::test_message_survives_the_worker_to_runner_hop": 0.7684,"tests/unit/pipelex/cogt/test_worker_error_enrichment.py::TestWorkerErrorEnrichment::test_extract_worker_enriches_error": 0.0396,@@ -13154,6 +13222,22 @@"tests/unit/pipelex/providers/anthropic/test_anthropic_structure_method.py::TestAnthropicStructureMethod::test_no_thinking_effort_keeps_the_forced_tool_choice": 1.4457,"tests/unit/pipelex/providers/anthropic/test_anthropic_structure_method.py::TestAnthropicStructureMethod::test_reasoning_tools_leaves_the_tool_choice_on_auto": 0.6873,"tests/unit/pipelex/providers/anthropic/test_anthropic_structure_method.py::TestAnthropicStructureMethod::test_tools_forces_the_response_tool": 0.6455,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_structured_clamp.py::TestAnthropicStructuredClamp::test_a_lowered_caller_max_tokens_is_logged_as_a_warning": 0.1516,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_structured_clamp.py::TestAnthropicStructuredClamp::test_a_lowered_model_default_is_logged_at_debug_level": 0.012,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_structured_clamp.py::TestAnthropicStructuredClamp::test_a_timeout_of_the_lowered_call_says_the_limit_was_lowered[False]": 0.0089,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_structured_clamp.py::TestAnthropicStructuredClamp::test_a_timeout_of_the_lowered_call_says_the_limit_was_lowered[True]": 0.0187,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_structured_clamp.py::TestAnthropicStructuredClamp::test_an_error_of_a_call_sent_as_given_carries_no_note": 0.1229,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_structured_clamp.py::TestAnthropicStructuredClamp::test_an_error_the_lowered_limit_cannot_explain_carries_no_note[authentication-failure1-False]": 0.0063,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_structured_clamp.py::TestAnthropicStructuredClamp::test_an_error_the_lowered_limit_cannot_explain_carries_no_note[authentication-failure1-True]": 0.0071,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_structured_clamp.py::TestAnthropicStructuredClamp::test_an_error_the_lowered_limit_cannot_explain_carries_no_note[lost connection-failure4-False]": 0.1806,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_structured_clamp.py::TestAnthropicStructuredClamp::test_an_error_the_lowered_limit_cannot_explain_carries_no_note[lost connection-failure4-True]": 0.0143,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_structured_clamp.py::TestAnthropicStructuredClamp::test_an_error_the_lowered_limit_cannot_explain_carries_no_note[overloaded-failure2-False]": 0.1239,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_structured_clamp.py::TestAnthropicStructuredClamp::test_an_error_the_lowered_limit_cannot_explain_carries_no_note[overloaded-failure2-True]": 0.0086,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_structured_clamp.py::TestAnthropicStructuredClamp::test_an_error_the_lowered_limit_cannot_explain_carries_no_note[rate limit-failure0-False]": 0.1607,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_structured_clamp.py::TestAnthropicStructuredClamp::test_an_error_the_lowered_limit_cannot_explain_carries_no_note[rate limit-failure0-True]": 0.0146,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_structured_clamp.py::TestAnthropicStructuredClamp::test_an_error_the_lowered_limit_cannot_explain_carries_no_note[server error-failure3-False]": 0.1511,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_structured_clamp.py::TestAnthropicStructuredClamp::test_an_error_the_lowered_limit_cannot_explain_carries_no_note[server error-failure3-True]": 0.0104,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_structured_clamp.py::TestAnthropicStructuredClamp::test_an_invalid_output_of_the_lowered_call_says_the_limit_was_lowered": 0.0166,"tests/unit/pipelex/providers/anthropic/test_anthropic_structured_thinking.py::TestAnthropicStructuredThinking::test_a_model_without_thinking_refuses_it_as_its_text_path_does[instructor/anthropic_reasoning_tools]": 0.0859,"tests/unit/pipelex/providers/anthropic/test_anthropic_structured_thinking.py::TestAnthropicStructuredThinking::test_a_model_without_thinking_refuses_it_as_its_text_path_does[instructor/anthropic_tools]": 0.0308,"tests/unit/pipelex/providers/anthropic/test_anthropic_structured_thinking.py::TestAnthropicStructuredThinking::test_a_reask_carries_the_signed_thinking_block_back[instructor/anthropic_reasoning_tools]": 0.0287,@@ -13167,6 +13251,11 @@"tests/unit/pipelex/providers/anthropic/test_anthropic_text_request.py::TestAnthropicTextRequest::test_a_model_refusing_temperature_omits_it_without_thinking": 0.2296,"tests/unit/pipelex/providers/anthropic/test_anthropic_text_request.py::TestAnthropicTextRequest::test_thinking_drops_the_temperature": 0.0102,"tests/unit/pipelex/providers/anthropic/test_anthropic_text_request.py::TestAnthropicTextRequest::test_without_thinking_the_temperature_is_sent": 0.7709,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_text_stop.py::TestAnthropicTextStop::test_a_normal_stop_returns_the_text_unchanged": 0.0022,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_text_stop.py::TestAnthropicTextStop::test_a_refusal_raises_the_refusal": 0.1302,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_text_stop.py::TestAnthropicTextStop::test_a_text_cut_at_max_tokens_raises_the_truncation[content0]": 0.0842,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_text_stop.py::TestAnthropicTextStop::test_a_text_cut_at_max_tokens_raises_the_truncation[content1]": 0.0035,+ "tests/unit/pipelex/providers/anthropic/test_anthropic_text_stop.py::TestAnthropicTextStop::test_a_text_cut_at_max_tokens_raises_the_truncation[content2]": 0.0076,"tests/unit/pipelex/providers/anthropic/test_anthropic_worker_error_handling.py::TestAnthropicWorkerErrorHandling::test_auth_error_is_configuration": 0.0764,"tests/unit/pipelex/providers/anthropic/test_anthropic_worker_error_handling.py::TestAnthropicWorkerErrorHandling::test_bad_request[content_policy-Your request was rejected due to content_policy_violation-content-safety filters]": 0.1068,"tests/unit/pipelex/providers/anthropic/test_anthropic_worker_error_handling.py::TestAnthropicWorkerErrorHandling::test_bad_request[generic_bad_request-Invalid parameter: max_tokens must be positive-content-None]": 0.0918,@@ -13282,6 +13371,12 @@"tests/unit/pipelex/providers/bedrock/test_bedrock_worker_semantic.py::TestBedrockWorkerSemantic::test_client_error_carries_semantic_user_action[UnknownAWSException-Mystery failure-500-transient-wait_and_retry]": 0.0496,"tests/unit/pipelex/providers/bedrock/test_bedrock_worker_semantic.py::TestBedrockWorkerSemantic::test_client_error_carries_semantic_user_action[ValidationException-Invalid request-400-content-change_input]": 0.0436,"tests/unit/pipelex/providers/bedrock/test_bedrock_worker_semantic.py::TestBedrockWorkerSemantic::test_resource_not_found_raises_llm_model_not_found_error": 0.1527,+ "tests/unit/pipelex/providers/bedrock/test_bedrock_worker_stop.py::TestBedrockWorkerStop::test_a_filtered_text_raises_the_refusal[content_filtered]": 0.0278,+ "tests/unit/pipelex/providers/bedrock/test_bedrock_worker_stop.py::TestBedrockWorkerStop::test_a_filtered_text_raises_the_refusal[guardrail_intervened]": 0.2876,+ "tests/unit/pipelex/providers/bedrock/test_bedrock_worker_stop.py::TestBedrockWorkerStop::test_a_normal_stop_returns_the_text_unchanged": 0.0308,+ "tests/unit/pipelex/providers/bedrock/test_bedrock_worker_stop.py::TestBedrockWorkerStop::test_a_normal_stop_with_no_text_is_an_empty_text_error": 0.1765,+ "tests/unit/pipelex/providers/bedrock/test_bedrock_worker_stop.py::TestBedrockWorkerStop::test_a_text_cut_at_max_tokens_raises_the_truncation[The contract binds the parties to deliver the goods by the end of]": 0.4466,+ "tests/unit/pipelex/providers/bedrock/test_bedrock_worker_stop.py::TestBedrockWorkerStop::test_a_text_cut_at_max_tokens_raises_the_truncation[]": 0.0116,"tests/unit/pipelex/providers/bedrock/test_extract_bedrock_metadata.py::TestExtractBedrockMetadata::test_access_denied_exception": 0.0004,"tests/unit/pipelex/providers/bedrock/test_extract_bedrock_metadata.py::TestExtractBedrockMetadata::test_full_response_metadata": 0.7652,"tests/unit/pipelex/providers/bedrock/test_extract_bedrock_metadata.py::TestExtractBedrockMetadata::test_invalid_retry_after_returns_none": 0.0003,@@ -13434,6 +13529,13 @@"tests/unit/pipelex/providers/google/test_google_llm_worker_text_request.py::TestGoogleLLMWorkerTextRequest::test_a_model_refusing_temperature_gets_none": 0.0201,"tests/unit/pipelex/providers/google/test_google_llm_worker_text_request.py::TestGoogleLLMWorkerTextRequest::test_an_adaptive_thinking_level_is_logged_as_its_wire_value": 0.1657,"tests/unit/pipelex/providers/google/test_google_llm_worker_text_request.py::TestGoogleLLMWorkerTextRequest::test_the_temperature_is_sent": 0.4261,+ "tests/unit/pipelex/providers/google/test_google_llm_worker_text_stop.py::TestGoogleLLMWorkerTextStop::test_a_blocked_prompt_raises_the_refusal_naming_its_reason[PROHIBITED_CONTENT]": 0.1408,+ "tests/unit/pipelex/providers/google/test_google_llm_worker_text_stop.py::TestGoogleLLMWorkerTextStop::test_a_blocked_prompt_raises_the_refusal_naming_its_reason[SAFETY]": 0.0225,+ "tests/unit/pipelex/providers/google/test_google_llm_worker_text_stop.py::TestGoogleLLMWorkerTextStop::test_a_candidate_cut_at_max_tokens_raises_the_truncation[The contract binds the parties to deliver the goods by the end of]": 0.1509,+ "tests/unit/pipelex/providers/google/test_google_llm_worker_text_stop.py::TestGoogleLLMWorkerTextStop::test_a_candidate_cut_at_max_tokens_raises_the_truncation[]": 0.0148,+ "tests/unit/pipelex/providers/google/test_google_llm_worker_text_stop.py::TestGoogleLLMWorkerTextStop::test_a_normal_stop_returns_the_text_unchanged": 0.0391,+ "tests/unit/pipelex/providers/google/test_google_llm_worker_text_stop.py::TestGoogleLLMWorkerTextStop::test_a_safety_stop_after_partial_text_raises_the_refusal": 0.0117,+ "tests/unit/pipelex/providers/google/test_google_llm_worker_text_stop.py::TestGoogleLLMWorkerTextStop::test_no_candidate_and_no_block_reason_keeps_the_generic_error": 0.0169,"tests/unit/pipelex/providers/google/test_google_reasoning.py::TestGoogleReasoning::test_adaptive_mode_effort_maps_to_thinking_level[high-HIGH]": 0.0529,"tests/unit/pipelex/providers/google/test_google_reasoning.py::TestGoogleReasoning::test_adaptive_mode_effort_maps_to_thinking_level[low-LOW]": 0.0552,"tests/unit/pipelex/providers/google/test_google_reasoning.py::TestGoogleReasoning::test_adaptive_mode_effort_maps_to_thinking_level[max-HIGH]": 0.0657,@@ -13703,6 +13805,11 @@"tests/unit/pipelex/providers/mistral/test_mistral_text_request.py::TestMistralTextRequest::test_a_model_refusing_temperature_omits_it": 0.3425,"tests/unit/pipelex/providers/mistral/test_mistral_text_request.py::TestMistralTextRequest::test_a_reasoning_effort_reaches_the_api_as_reasoning_effort": 1.78,"tests/unit/pipelex/providers/mistral/test_mistral_text_request.py::TestMistralTextRequest::test_without_a_reasoning_setting_no_effort_is_sent": 2.5694,+ "tests/unit/pipelex/providers/mistral/test_mistral_text_stop.py::TestMistralTextStop::test_a_filter_value_raises_the_refusal": 0.0115,+ "tests/unit/pipelex/providers/mistral/test_mistral_text_stop.py::TestMistralTextStop::test_a_normal_stop_returns_the_text_unchanged": 0.018,+ "tests/unit/pipelex/providers/mistral/test_mistral_text_stop.py::TestMistralTextStop::test_a_text_cut_at_its_length_raises_the_truncation[length--max_tokens]": 0.1746,+ "tests/unit/pipelex/providers/mistral/test_mistral_text_stop.py::TestMistralTextStop::test_a_text_cut_at_its_length_raises_the_truncation[length-The contract binds the parties to deliver the goods by the end of-max_tokens]": 0.1245,+ "tests/unit/pipelex/providers/mistral/test_mistral_text_stop.py::TestMistralTextStop::test_a_text_cut_at_its_length_raises_the_truncation[model_length-The contract binds the parties to deliver the goods by the end of-context_window]": 0.0101,"tests/unit/pipelex/providers/mistral/test_mistral_worker_error_handling.py::TestMistralWorkerErrorHandling::test_extract_error_report_includes_category": 0.0911,"tests/unit/pipelex/providers/mistral/test_mistral_worker_error_handling.py::TestMistralWorkerErrorHandling::test_extract_worker_error_classification[auth_error_401-401-Invalid API key provided-configuration-None]": 0.0023,"tests/unit/pipelex/providers/mistral/test_mistral_worker_error_handling.py::TestMistralWorkerErrorHandling::test_extract_worker_error_classification[bad_request_content_policy_400-400-Your request was rejected due to content_policy_violation-content-safety filters]": 0.0034,@@ -13743,6 +13850,14 @@"tests/unit/pipelex/providers/openai/test_extract_openai_metadata.py::TestExtractOpenAIMetadata::test_retry_after_seconds_is_none_for_unparseable_value": 0.0013,"tests/unit/pipelex/providers/openai/test_extract_openai_metadata.py::TestExtractOpenAIMetadata::test_retry_after_seconds_is_none_when_header_absent": 0.0005,"tests/unit/pipelex/providers/openai/test_extract_openai_metadata.py::TestExtractOpenAIMetadata::test_retry_after_seconds_parses_future_http_date": 0.0233,+ "tests/unit/pipelex/providers/openai/test_openai_completions_gateway_stop.py::TestOpenAICompletionsGatewayStop::test_a_passed_through_normal_stop_returns_the_text": 0.4965,+ "tests/unit/pipelex/providers/openai/test_openai_completions_gateway_stop.py::TestOpenAICompletionsGatewayStop::test_a_passed_through_refusal_or_filter_raises_the_refusal[bedrock content filter-content_filtered-]": 0.8905,+ "tests/unit/pipelex/providers/openai/test_openai_completions_gateway_stop.py::TestOpenAICompletionsGatewayStop::test_a_passed_through_refusal_or_filter_raises_the_refusal[claude refusal-refusal-]": 0.0196,+ "tests/unit/pipelex/providers/openai/test_openai_completions_gateway_stop.py::TestOpenAICompletionsGatewayStop::test_a_passed_through_refusal_or_filter_raises_the_refusal[gemini safety stop after partial text-SAFETY-The contract binds the parties to]": 0.0149,+ "tests/unit/pipelex/providers/openai/test_openai_completions_gateway_stop.py::TestOpenAICompletionsGatewayStop::test_a_passed_through_truncation_raises_the_truncation[claude cut at max_tokens-max_tokens-The contract binds the parties to]": 0.6602,+ "tests/unit/pipelex/providers/openai/test_openai_completions_gateway_stop.py::TestOpenAICompletionsGatewayStop::test_a_passed_through_truncation_raises_the_truncation[claude cut with nothing written-max_tokens-]": 0.0102,+ "tests/unit/pipelex/providers/openai/test_openai_completions_gateway_stop.py::TestOpenAICompletionsGatewayStop::test_a_passed_through_truncation_raises_the_truncation[gemini cut at max_tokens-MAX_TOKENS-The contract binds the parties to]": 0.4958,+ "tests/unit/pipelex/providers/openai/test_openai_completions_gateway_stop.py::TestOpenAICompletionsGatewayStop::test_a_passed_through_truncation_raises_the_truncation[strict mode cut-length-The contract binds the parties to]": 0.0939,"tests/unit/pipelex/providers/openai/test_openai_completions_img_gen_worker_error_handling.py::TestOpenAICompletionsImgGenWorkerErrorHandling::test_auth_error_is_configuration_check_credentials": 0.0466,"tests/unit/pipelex/providers/openai/test_openai_completions_img_gen_worker_error_handling.py::TestOpenAICompletionsImgGenWorkerErrorHandling::test_connection_error_is_transient": 0.0578,"tests/unit/pipelex/providers/openai/test_openai_completions_img_gen_worker_error_handling.py::TestOpenAICompletionsImgGenWorkerErrorHandling::test_content_policy_bad_request_is_content_change_input": 0.0615,@@ -13777,6 +13892,13 @@"tests/unit/pipelex/providers/openai/test_openai_completions_reasoning.py::TestOpenAICompletionsReasoning::test_reasoning_effort_maps_correctly[xhigh-xhigh]": 0.0433,"tests/unit/pipelex/providers/openai/test_openai_completions_reasoning.py::TestOpenAICompletionsReasoning::test_thinking_mode_adaptive_raises_capability_error": 0.0012,"tests/unit/pipelex/providers/openai/test_openai_completions_reasoning.py::TestOpenAICompletionsReasoning::test_thinking_mode_none_raises_capability_error": 0.0018,+ "tests/unit/pipelex/providers/openai/test_openai_completions_retries_kept.py::TestOpenAICompletionsRetriesKept::test_a_rate_limited_text_call_stays_retryable": 0.6738,+ "tests/unit/pipelex/providers/openai/test_openai_completions_retries_kept.py::TestOpenAICompletionsRetriesKept::test_a_structured_output_failing_validation_is_retried_as_configured": 0.6895,+ "tests/unit/pipelex/providers/openai/test_openai_completions_retries_kept.py::TestOpenAICompletionsRetriesKept::test_the_configured_count_bounds_the_retries": 0.0189,+ "tests/unit/pipelex/providers/openai/test_openai_completions_stop.py::TestOpenAICompletionsStop::test_a_filtered_text_raises_the_refusal": 0.0111,+ "tests/unit/pipelex/providers/openai/test_openai_completions_stop.py::TestOpenAICompletionsStop::test_a_normal_stop_returns_the_text_unchanged": 0.4579,+ "tests/unit/pipelex/providers/openai/test_openai_completions_stop.py::TestOpenAICompletionsStop::test_a_text_cut_at_its_length_raises_the_truncation[The contract binds the parties to deliver the goods by the end of]": 0.1293,+ "tests/unit/pipelex/providers/openai/test_openai_completions_stop.py::TestOpenAICompletionsStop::test_a_text_cut_at_its_length_raises_the_truncation[]": 0.8821,"tests/unit/pipelex/providers/openai/test_openai_completions_structured_request.py::TestOpenAICompletionsStructuredRequest::test_a_model_refusing_temperature_still_omits_it_without_reasoning": 4.7145,"tests/unit/pipelex/providers/openai/test_openai_completions_structured_request.py::TestOpenAICompletionsStructuredRequest::test_a_model_without_reasoning_refuses_it_as_its_text_path_does": 8.1196,"tests/unit/pipelex/providers/openai/test_openai_completions_structured_request.py::TestOpenAICompletionsStructuredRequest::test_a_reasoning_effort_reaches_the_api_and_drops_the_temperature": 2.8207,@@ -13826,6 +13948,11 @@"tests/unit/pipelex/providers/openai/test_openai_responses_reasoning.py::TestOpenAIResponsesReasoning::test_reasoning_effort_maps_correctly[xhigh-xhigh]": 0.0649,"tests/unit/pipelex/providers/openai/test_openai_responses_reasoning.py::TestOpenAIResponsesReasoning::test_thinking_mode_adaptive_raises_capability_error": 0.001,"tests/unit/pipelex/providers/openai/test_openai_responses_reasoning.py::TestOpenAIResponsesReasoning::test_thinking_mode_none_raises_capability_error": 0.0043,+ "tests/unit/pipelex/providers/openai/test_openai_responses_stop.py::TestOpenAIResponsesStop::test_a_completed_answer_returns_the_text_unchanged": 0.5585,+ "tests/unit/pipelex/providers/openai/test_openai_responses_stop.py::TestOpenAIResponsesStop::test_a_filtered_answer_raises_the_refusal": 0.5858,+ "tests/unit/pipelex/providers/openai/test_openai_responses_stop.py::TestOpenAIResponsesStop::test_an_answer_cut_at_its_output_limit_raises_the_truncation[The contract binds the parties to deliver the goods by the end of]": 0.2059,+ "tests/unit/pipelex/providers/openai/test_openai_responses_stop.py::TestOpenAIResponsesStop::test_an_answer_cut_at_its_output_limit_raises_the_truncation[]": 0.7086,+ "tests/unit/pipelex/providers/openai/test_openai_responses_stop.py::TestOpenAIResponsesStop::test_an_incomplete_answer_with_no_details_raises_the_truncation": 0.0188,"tests/unit/pipelex/providers/openai/test_openai_responses_structured_request.py::TestOpenAIResponsesStructuredRequest::test_a_model_refusing_temperature_omits_it_without_reasoning": 6.8696,"tests/unit/pipelex/providers/openai/test_openai_responses_structured_request.py::TestOpenAIResponsesStructuredRequest::test_a_model_without_reasoning_refuses_it_as_its_text_path_does": 0.0417,"tests/unit/pipelex/providers/openai/test_openai_responses_structured_request.py::TestOpenAIResponsesStructuredRequest::test_a_reasoning_budget_is_refused_as_its_text_path_does": 2.4075,diff --git a/CHANGELOG.md b/CHANGELOG.mdindex 810c573da..0e784c5f1 100644--- a/CHANGELOG.md+++ b/CHANGELOG.md@@ -1,10 +1,10 @@# Changelog-## [Unreleased]+## [v0.83.0] - 2026-10-11### Added-- **`LLMCompletionTruncatedError` and `LLMCompletionRefusedError`**: two subclasses of `LLMCompletionError` report a text completion the model did not finish, the first one cut at its output limit or context window, the second one refused by the model or stopped by a safety filter. Both are content errors in the `input` domain, not retryable, answered with HTTP 422 and caller-facing under STRICT disclosure, with a `change_input` next step; the message is one sentence naming the model, the pipe, the provider's stop value and, for a truncation, the output tokens used and the `max_tokens` sent, and never carries the partial text. A truncation carries the limit it hit as `truncation_limit`, `max_tokens` or `context_window`, a stop at the context window advises a shorter input, a lower reasoning effort or a model with a larger window rather than a higher `max_tokens`, and a cut when the request sent no `max_tokens` advises setting the pipe's up to the model's limit.+- **`LLMCompletionTruncatedError` and `LLMCompletionRefusedError`**: two subclasses of `LLMCompletionError`, the first for a text completion cut at its output limit or its context window, the second for one the model declined to write or a safety filter stopped. Both are content errors in the `input` domain, not retryable, answered with HTTP 422 and caller-facing under STRICT disclosure, and carry `model_handle`, `pipe_code` and the provider's `stop_reason`, a truncation also carrying the limit it hit as `truncation_limit`, `max_tokens` or `context_window`, the `output_tokens` used and the `max_tokens` sent; the message names the model, the pipe and those values, never the partial text. Each ends with a `change_input` next step fitted to what stopped the text: revise the prompt or the input after a refusal, raise the pipe's `max_tokens` or shorten its input at the output limit, set the pipe's `max_tokens` up to the model's limit when the request sent none, and shorten the input, lower the reasoning effort or choose a model with a larger window, never a higher `max_tokens`, at the context window.### Changed@@ -13,7 +13,7 @@### Fixed-- **A truncated, refused or filtered text no longer passes for a result (Breaking)**: every LLM worker, OpenAI chat completions with a gateway, Portkey or OpenRouter in front of it, the OpenAI Responses API, Anthropic, Google, Bedrock and Mistral, now reads its response's stop value before it returns a text, so a text cut at `max_tokens` or refused fails its pipe with `LLMCompletionTruncatedError` or `LLMCompletionRefusedError` where it used to come back as a successful output, empty or partial, its billed usage still reported and recorded on its span. Each provider's vocabulary is read, including a gateway's pass-through of another provider's value such as `max_tokens` or `content_filtered`, an unknown value being logged at warning level and taken as normal, except that an `incomplete` Responses API answer is a truncation whatever its reason; a Gemini prompt blocked before any candidate fails with `LLMCompletionRefusedError` naming its block reason, and structured generation keeps its configured re-asks. A Bedrock answer holding no text block fails with an `LLMCompletionError` instead of an `IndexError`.+- **A truncated, refused or filtered text no longer passes for a result (Breaking)**: every LLM worker, OpenAI chat completions whether direct or behind a gateway such as Portkey or OpenRouter, the OpenAI Responses API, Anthropic, Google, Bedrock and Mistral, now reads its response's stop value before it returns a text, so a text that stopped short fails its pipe with `LLMCompletionTruncatedError` or `LLMCompletionRefusedError` where it used to come back as a successful output, empty or partial, its billed usage still reported and recorded on its span. Each provider's vocabulary is read, including a gateway's pass-through of another provider's value such as `max_tokens` or `content_filtered`, an unknown value being logged at warning level and taken as normal, except that an `incomplete` Responses API answer is a truncation whatever its reason; a Gemini prompt blocked before any candidate is a refusal, and structured generation keeps its configured re-asks. A Bedrock answer is read for every text block it holds, so one holding none fails with an `LLMCompletionError` where it used to raise an `IndexError` when its message was empty, or a `KeyError` when its first block was not text.## [v0.82.0] - 2026-10-10diff --git a/docs/api-server/openapi/pipelex-api.openapi.yaml b/docs/api-server/openapi/pipelex-api.openapi.yamlindex 7ea6585de..547545521 100644--- a/docs/api-server/openapi/pipelex-api.openapi.yaml+++ b/docs/api-server/openapi/pipelex-api.openapi.yaml@@ -12,7 +12,7 @@ info:license:name: Elastic License 2.0identifier: Elastic-2.0- version: 0.82.0+ version: 0.83.0paths:/health:get:diff --git a/pyproject.toml b/pyproject.tomlindex 9ded2c657..0d19d4f5a 100644--- a/pyproject.toml+++ b/pyproject.toml@@ -1,6 +1,6 @@[project]name = "pipelex"-version = "0.82.0"+version = "0.83.0"description = "Execute composable AI methods declared in the MTHDS open standard"authors = [{ name = "Evotis S.A.S.", email = "[email protected]" }]maintainers = [{ name = "Pipelex staff", email = "[email protected]" }]diff --git a/uv.lock b/uv.lockindex 01abf7f0e..d3aca5392 100644--- a/uv.lock+++ b/uv.lock@@ -3530,7 +3530,7 @@ wheels = [[[package]]name = "pipelex"-version = "0.82.0"+version = "0.83.0"source = { editable = "." }dependencies = [{ name = "aiofiles" },