repo stringclasses 454
values | file_path stringlengths 5 201 | extension stringclasses 1
value | content stringlengths 8 509k | num_lines int64 3 16.9k | size_bytes int64 8 511k |
|---|---|---|---|---|---|
cognee | cognee/infrastructure/data/exceptions/exceptions.py | .py | from fastapi import status
from cognee.exceptions import (
CogneeValidationError,
)
class KeywordExtractionError(CogneeValidationError):
"""
Raised when a provided value is syntactically valid but semantically unacceptable
for the given operation.
Example:
- Passing an empty string to a ... | 24 | 644 |
cognee | cognee/infrastructure/data/exceptions/__init__.py | .py | """
Custom exceptions for the Cognee API.
This module defines a set of exceptions for handling various data errors
"""
from .exceptions import KeywordExtractionError
| 8 | 168 |
cognee | cognee/infrastructure/data/chunking/HaystackChunkEngine.py | .py | from collections.abc import Iterable
from cognee.shared.data_models import ChunkStrategy
class HaystackChunkEngine:
"""
Manage chunking of source data using specified strategies and parameters.
The class provides functionality to process source data into manageable chunks according
to defined strate... | 27 | 903 |
cognee | cognee/infrastructure/data/chunking/LangchainChunkingEngine.py | .py | from __future__ import annotations
import re
from collections.abc import Iterable
from typing import Any
from cognee.infrastructure.data.chunking.DefaultChunkEngine import DefaultChunkEngine
from cognee.shared.data_models import ChunkStrategy
class LangchainChunkEngine:
"""
Handles chunking of data using sp... | 159 | 5,594 |
cognee | cognee/infrastructure/data/chunking/get_chunking_engine.py | .py | from typing import Any
from cognee.infrastructure.data.chunking.DefaultChunkEngine import DefaultChunkEngine
from cognee.infrastructure.data.chunking.HaystackChunkEngine import HaystackChunkEngine
from cognee.infrastructure.data.chunking.LangchainChunkingEngine import LangchainChunkEngine
from .config import get_chun... | 22 | 739 |
cognee | cognee/infrastructure/data/chunking/DefaultChunkEngine.py | .py | """Chunking strategies for splitting text into smaller parts."""
from __future__ import annotations
import re
from collections.abc import Iterable
from typing import Any
from cognee.shared.data_models import ChunkStrategy
# /Users/vasa/Projects/cognee/cognee/infrastructure/data/chunking/DefaultChunkEngine.py
clas... | 219 | 7,892 |
cognee | cognee/infrastructure/data/chunking/create_chunking_engine.py | .py | from typing import Any
from cognee.infrastructure.data.chunking.DefaultChunkEngine import DefaultChunkEngine
from cognee.infrastructure.data.chunking.HaystackChunkEngine import HaystackChunkEngine
from cognee.infrastructure.data.chunking.LangchainChunkingEngine import LangchainChunkEngine
from cognee.shared.data_model... | 71 | 2,821 |
cognee | cognee/infrastructure/data/chunking/config.py | .py | from functools import lru_cache
from typing import Any
from pydantic_settings import BaseSettings, SettingsConfigDict
from cognee.infrastructure.data.chunking.DefaultChunkEngine import DefaultChunkEngine
from cognee.shared.data_models import ChunkEngine, ChunkStrategy
class ChunkConfig(BaseSettings):
"""
Ma... | 56 | 1,725 |
cognee | cognee/infrastructure/locks/dataset_lock.py | .py | """Per-dataset lock primitives — in-process asyncio registry.
Serializes operations that mutate the same dataset — pipeline runs
(``add``/``cognify``/``memify``) and delete operations — while letting
different datasets proceed in parallel. Both acquire the lock from the
same registry, so a delete waits for an in-fligh... | 67 | 2,480 |
cognee | cognee/infrastructure/locks/__init__.py | .py | from .dataset_lock import dataset_lock, get_dataset_lock, held_datasets
from .session_lock import (
release_improve_lock,
session_lock,
session_turn_lock,
try_acquire_improve_lock,
)
__all__ = [
"dataset_lock",
"get_dataset_lock",
"held_datasets",
"release_improve_lock",
"session_lo... | 18 | 384 |
cognee | cognee/infrastructure/locks/session_lock.py | .py | """Per-session lock primitives — in-process asyncio registry.
Three primitives:
* ``session_lock(session_id, op)`` — async context manager that
serializes concurrent tasks on the same ``(session_id, op)`` key.
Used for short read-modify-write flows (``update_qa``,
``add_feedback``, ``delete_qa``).
* ``session_... | 138 | 4,771 |
cognee | cognee/eval_framework/run_beam_eval.py | .py | """Run BEAM benchmark evaluation on a single conversation.
Usage:
uv run python cognee/eval_framework/run_beam_eval.py
"""
import asyncio
from cognee.shared.logging_utils import get_logger
from cognee.eval_framework.eval_config import EvalConfig
from cognee.eval_framework.corpus_builder.run_corpus_builder import... | 83 | 2,704 |
cognee | cognee/eval_framework/modal_eval_dashboard.py | .py | import os
import json
import shlex
import subprocess
import modal
import streamlit as st
# ----------------------------------------------------------------------------
# Volume and Image Setup
# ----------------------------------------------------------------------------
metrics_volume = modal.Volume.from_name("evalu... | 112 | 3,585 |
cognee | cognee/eval_framework/runner.py | .py | """One-command runner for the cognee evaluation harness.
This module puts a thin, testable surface in front of the existing eval pipeline
(corpus -> answers -> evaluation -> dashboard). It exposes:
- ``run_eval(config) -> EvalResult``: a programmatic entry point that runs the
full pipeline for a single deterministi... | 325 | 11,544 |
cognee | cognee/eval_framework/__main__.py | .py | """Module entry point: ``python -m cognee.eval_framework [options]``.
Thin wrapper around :func:`cognee.eval_framework.runner.run_eval` that shares its
argument surface with the ``cognee eval`` CLI command.
"""
import argparse
import asyncio
from cognee.eval_framework.runner import (
add_eval_arguments,
conf... | 36 | 865 |
cognee | cognee/eval_framework/eval_config.py | .py | from functools import lru_cache
from pydantic_settings import BaseSettings, SettingsConfigDict
from typing import List, Optional
class EvalConfig(BaseSettings):
# Corpus builder params
building_corpus_from_scratch: bool = True
number_of_samples_in_corpus: int = 1
benchmark: str = "Dummy" # Options: '... | 92 | 3,691 |
cognee | cognee/eval_framework/run_eval.py | .py | """Legacy script entry point for the eval harness.
Kept for backward compatibility. New code should prefer the one-command runner:
from cognee.eval_framework.runner import run_eval
result = await run_eval(EvalConfig())
or the CLI: ``cognee eval ...`` / ``python -m cognee.eval_framework``.
"""
import asynci... | 34 | 852 |
cognee | cognee/eval_framework/modal_run_eval.py | .py | import modal
import os
import asyncio
import datetime
import json
from cognee.shared.logging_utils import get_logger
from cognee.eval_framework.eval_config import EvalConfig
from cognee.eval_framework.corpus_builder.run_corpus_builder import run_corpus_builder
from cognee.eval_framework.answer_generation.run_question_a... | 132 | 4,458 |
cognee | cognee/eval_framework/metrics_dashboard.py | .py | import html
import json
import plotly.graph_objects as go
from typing import Dict, List, Tuple
from collections import defaultdict
metrics_fields = {
"contextual_relevancy": ["question", "retrieval_context"],
"context_coverage": ["question", "retrieval_context", "golden_context"],
}
default_metrics_fields = ["... | 173 | 5,914 |
cognee | cognee/eval_framework/benchmark_adapters/base_benchmark_adapter.py | .py | from abc import ABC, abstractmethod
from typing import List, Optional, Any, Union, Tuple
import os
import json
from cognee.shared.logging_utils import get_logger
logger = get_logger()
class BaseBenchmarkAdapter(ABC):
async def prepare_corpus(self) -> None:
return None
def _filter_instances(
... | 59 | 2,306 |
cognee | cognee/eval_framework/benchmark_adapters/hotpot_qa_adapter.py | .py | import os
import json
import random
from typing import Optional, Any, List, Union, Tuple
from cognee.eval_framework.benchmark_adapters.base_benchmark_adapter import BaseBenchmarkAdapter
class HotpotQAAdapter(BaseBenchmarkAdapter):
dataset_info = {
"filename": "hotpot_benchmark.json",
"url": "http:... | 107 | 4,331 |
cognee | cognee/eval_framework/benchmark_adapters/dummy_adapter.py | .py | from typing import Optional, Any, List, Union, Tuple
from cognee.eval_framework.benchmark_adapters.base_benchmark_adapter import BaseBenchmarkAdapter
class DummyAdapter(BaseBenchmarkAdapter):
def load_corpus(
self,
limit: Optional[int] = None,
seed: int = 42,
load_golden_context: ... | 33 | 1,167 |
cognee | cognee/eval_framework/benchmark_adapters/logistics_system_adapter.py | .py | from __future__ import annotations
import json
import random
from pathlib import Path
from typing import Any, List, Optional, Tuple, Union
from cognee.eval_framework.benchmark_adapters.base_benchmark_adapter import (
BaseBenchmarkAdapter,
)
from cognee.eval_framework.benchmark_adapters.logistics_system_utils.corp... | 229 | 8,659 |
cognee | cognee/eval_framework/benchmark_adapters/benchmark_adapters.py | .py | from enum import Enum
from typing import Type
from cognee.eval_framework.benchmark_adapters.hotpot_qa_adapter import HotpotQAAdapter
from cognee.eval_framework.benchmark_adapters.logistics_system_adapter import (
LogisticsSystemAdapter,
)
from cognee.eval_framework.benchmark_adapters.musique_adapter import Musique... | 30 | 1,153 |
cognee | cognee/eval_framework/benchmark_adapters/twowikimultihop_adapter.py | .py | from typing import Any
from cognee.eval_framework.benchmark_adapters.hotpot_qa_adapter import HotpotQAAdapter
class TwoWikiMultihopAdapter(HotpotQAAdapter):
dataset_info = {
"filename": "2wikimultihop_dev.json",
"url": "https://huggingface.co/datasets/voidful/2WikiMultihopQA/resolve/main/dev.json"... | 25 | 913 |
cognee | cognee/eval_framework/benchmark_adapters/beam_adapter.py | .py | """BEAM benchmark adapter — loads synthetic long-context conversations.
Dataset: https://huggingface.co/datasets/Mohammadta/BEAM
Paper: "Beyond a Million Tokens: Benchmarking Long-Term Memory in LLMs"
Each conversation contains multi-session batches with 20 probing questions
across 10 skill categories (information_ex... | 258 | 9,198 |
cognee | cognee/eval_framework/benchmark_adapters/musique_adapter.py | .py | import os
import json
import random
from typing import Optional, Any, List, Union, Tuple
import zipfile
from cognee.eval_framework.benchmark_adapters.base_benchmark_adapter import BaseBenchmarkAdapter
class MusiqueQAAdapter(BaseBenchmarkAdapter):
"""Adapter for the Musique QA dataset with local file loading and ... | 142 | 5,613 |
cognee | cognee/eval_framework/benchmark_adapters/logistics_system_utils/ontology.py | .py | from __future__ import annotations
import random
import json
from pathlib import Path
from cognee.eval_framework.benchmark_adapters.logistics_system_utils.entities.package import Package
from cognee.eval_framework.benchmark_adapters.logistics_system_utils.entities.retailer import (
retailer_possible_values,
)
from... | 228 | 7,874 |
cognee | cognee/eval_framework/benchmark_adapters/logistics_system_utils/rule_engine.py | .py | from __future__ import annotations
from math import inf
from typing import Any
from cognee.eval_framework.benchmark_adapters.logistics_system_utils.entities.delivery_models import (
DeliveryPlan,
RejectedOption,
TransportEvaluation,
)
from cognee.eval_framework.benchmark_adapters.logistics_system_utils.en... | 730 | 28,317 |
cognee | cognee/eval_framework/benchmark_adapters/logistics_system_utils/entities/transportation.py | .py | from __future__ import annotations
from dataclasses import dataclass
from cognee.eval_framework.benchmark_adapters.logistics_system_utils.entities.enums import (
Region,
ShippingRange,
TransportMode,
)
transport_possible_values = {
"region": [region.label for region in Region],
"temperature_cont... | 44 | 1,319 |
cognee | cognee/eval_framework/benchmark_adapters/logistics_system_utils/entities/user.py | .py | from __future__ import annotations
from dataclasses import dataclass
from cognee.eval_framework.benchmark_adapters.logistics_system_utils.entities.enums import (
Region,
ShippingRange,
UserTier,
)
user_possible_values = {
"weekend_delivery_eligible": [True, False],
"user_tier": [tier.label for t... | 47 | 1,215 |
cognee | cognee/eval_framework/benchmark_adapters/logistics_system_utils/entities/enums.py | .py | from enum import Enum
class UserTier(Enum):
STANDARD = ("standard", 0, 1)
BUSINESS = ("business", 1, 2)
ENTERPRISE = ("enterprise", 2, 3)
@property
def label(self) -> str:
return self.value[0]
@property
def priority_bonus(self) -> int:
return self.value[1]
@property
... | 238 | 6,247 |
cognee | cognee/eval_framework/benchmark_adapters/logistics_system_utils/entities/package.py | .py | from __future__ import annotations
from dataclasses import dataclass
from cognee.eval_framework.benchmark_adapters.logistics_system_utils.entities.enums import (
DeliveryPriority,
PackageCategory,
PackageState,
ShippingRange,
)
package_possible_values = {
"weight_kg": [1.0, 5.0, 25.0, 150.0, 900... | 134 | 4,340 |
cognee | cognee/eval_framework/benchmark_adapters/logistics_system_utils/entities/delivery_models.py | .py | from __future__ import annotations
import json
from dataclasses import dataclass, field
from cognee.eval_framework.benchmark_adapters.logistics_system_utils.entities.enums import (
TransportMode,
)
from cognee.eval_framework.benchmark_adapters.logistics_system_utils.entities.package import Package
from cognee.eva... | 165 | 6,337 |
cognee | cognee/eval_framework/benchmark_adapters/logistics_system_utils/entities/retailer.py | .py | from __future__ import annotations
from dataclasses import dataclass
from cognee.eval_framework.benchmark_adapters.logistics_system_utils.entities.enums import (
Region,
ShippingRange,
)
retailer_possible_values = {
"region": [region.label for region in Region],
"shipping_range": [shipping_range.lab... | 49 | 1,285 |
cognee | cognee/eval_framework/benchmark_adapters/logistics_system_utils/entities/post_office.py | .py | from __future__ import annotations
from dataclasses import dataclass
from cognee.eval_framework.benchmark_adapters.logistics_system_utils.entities.enums import (
PostOfficeType,
Region,
ShippingRange,
)
post_office_possible_values = {
"office_type": [office_type.label for office_type in PostOfficeTy... | 51 | 1,465 |
cognee | cognee/eval_framework/benchmark_adapters/logistics_system_utils/utils/utils.py | .py | from __future__ import annotations
import json
import re
from pathlib import Path
from cognee.eval_framework.benchmark_adapters.logistics_system_utils.entities.package import Package
from cognee.eval_framework.benchmark_adapters.logistics_system_utils.entities.post_office import (
PostOffice,
)
from cognee.eval_f... | 352 | 14,265 |
cognee | cognee/eval_framework/benchmark_adapters/logistics_system_utils/utils/world_creation_utils.py | .py | import random
from cognee.eval_framework.benchmark_adapters.logistics_system_utils.entities.enums import (
DeliveryPriority,
PackageCategory,
PackageState,
PostOfficeType,
Region,
ShippingRange,
TransportMode,
)
from cognee.eval_framework.benchmark_adapters.logistics_system_utils.entities.p... | 799 | 27,411 |
cognee | cognee/eval_framework/benchmark_adapters/logistics_system_utils/corpus_generator/narrativize_corpus.py | .py | from __future__ import annotations
import asyncio
import os
from pathlib import Path
from pydantic import BaseModel
from cognee.eval_framework.benchmark_adapters.logistics_system_utils.utils.utils import (
_entity_entries,
_format_packages,
_safe_filename,
load_world,
)
from cognee.eval_framework.ben... | 240 | 9,276 |
cognee | cognee/eval_framework/reporting/io.py | .py | import json
from pathlib import Path
from typing import Any
def write_json(path: str, payload: Any) -> None:
Path(path).parent.mkdir(parents=True, exist_ok=True)
with open(path, "w", encoding="utf-8") as handle:
json.dump(payload, handle, ensure_ascii=False, indent=2)
def read_json(path: str) -> Any... | 15 | 409 |
cognee | cognee/eval_framework/analysis/dashboard_generator.py | .py | import html
import json
import plotly.graph_objects as go
from typing import Dict, List, Tuple
from collections import defaultdict
def create_distribution_plots(metrics_data: Dict[str, List[float]]) -> List[str]:
"""Create distribution histogram plots for each metric."""
figures = []
for metric, scores in... | 175 | 5,903 |
cognee | cognee/eval_framework/analysis/metrics_calculator.py | .py | import json
from collections import defaultdict
import numpy as np
from typing import Dict, List, Tuple
def bootstrap_ci(scores, num_samples=10000, confidence_level=0.95):
"""Calculate bootstrap confidence intervals for a list of scores."""
means = []
n = len(scores)
for _ in range(num_samples):
... | 95 | 3,392 |
cognee | cognee/eval_framework/sweeps/retriever_sweep_runner.py | .py | """Generic multi-strategy answer + scoring sweep."""
from __future__ import annotations
import asyncio
import re
from copy import deepcopy
from dataclasses import dataclass, field
from pathlib import Path
from typing import Any, Awaitable, Callable, Optional
from cognee.eval_framework.answer_generation.question_type... | 381 | 12,451 |
cognee | cognee/eval_framework/beam/session_io.py | .py | """Scale-agnostic ingestion-ready BEAM session-JSON contract.
One implementation of the read/parse/write primitives that both the local and Modal
ingestion paths need for BEAM's JSON-list session files, previously duplicated (with a
``_``-prefixed naming convention on the Modal side) between
``beam_ingest_conversation... | 78 | 3,000 |
cognee | cognee/eval_framework/beam/local_ingest.py | .py | """Sequentially ingest BEAM JSON-list sessions with parallel session distillation.
Given one conversation folder created by the preprocessing stage, this script ingests each
session/batch file one at a time:
1. write a limited copy of the session file into a run folder,
2. parse the known user/assistant turns for ses... | 608 | 21,140 |
cognee | cognee/eval_framework/beam/preprocessing/loaders.py | .py | """BEAM / BEAM-10M dataset loading, reworked to load once per run.
``cognee.eval_framework.benchmark_adapters.beam_adapter``/``beam_10m_adapter`` reload the
HuggingFace dataset on every call (``load_beam_row``/``load_beam_10m_row``), which is fine for
loading a single conversation for an eval run but wasteful for prep... | 89 | 3,192 |
cognee | cognee/eval_framework/beam/preprocessing/preprocess.py | .py | """Preprocess BEAM/BEAM-10M conversations into audited + ingestion-ready JSON documents.
Usage example:
HF_HUB_OFFLINE=1 HF_DATASETS_OFFLINE=1 uv run python \
-m cognee.eval_framework.beam.preprocessing.preprocess \
--dataset both --execute-compressions --output-dir temp/beam_preprocessed_documents... | 1,141 | 40,312 |
cognee | cognee/eval_framework/beam/preprocessing/conversation_preprocessing.py | .py | """Shared BEAM-aware conversation preprocessing helpers."""
from __future__ import annotations
from dataclasses import dataclass
from functools import lru_cache
import re
_TURN_START = re.compile(
r"^(?:\[(?P<time_anchor>.*?)\]\s*)?(?P<role>User|Assistant):\s?(?P<content>.*)$"
)
_SESSION_MARKER = re.compile(r"^-... | 955 | 30,422 |
cognee | cognee/eval_framework/beam/preprocessing/compression.py | .py | from __future__ import annotations
import argparse
import asyncio
import hashlib
import json
from collections import Counter
from collections.abc import Callable, Iterable
from dataclasses import dataclass
from typing import Any
from pydantic import BaseModel, Field
from cognee.eval_framework.beam.preprocessing.conv... | 572 | 19,797 |
cognee | cognee/eval_framework/beam/eval/beam_eval_adapter.py | .py | import asyncio
import os
from dataclasses import dataclass
from typing import Any, Dict, List, Optional
from cognee.eval_framework.evaluation.base_eval_adapter import BaseEvalAdapter
from cognee.eval_framework.beam.eval.metrics.beam_rubric import BEAMRubricMetric
from cognee.eval_framework.beam.eval.metrics.kendall_ta... | 121 | 4,207 |
cognee | cognee/eval_framework/beam/eval/aggregate_cross_run.py | .py | """Aggregate BEAM metrics across repeated runs for one retriever."""
import argparse
import json
import re
import statistics
from collections import Counter, defaultdict
from pathlib import Path
from cognee.eval_framework.analysis.metrics_calculator import bootstrap_ci
ARTIFACT_PREFIX = "beam_existing_ingestion"
MET... | 181 | 5,914 |
cognee | cognee/eval_framework/beam/eval/registry.py | .py | from dataclasses import dataclass, field
from typing import Any, Literal, Optional
from cognee.modules.retrieval.completion_retriever import CompletionRetriever
from cognee.modules.retrieval.graph_completion_context_extension_retriever import (
GraphCompletionContextExtensionRetriever,
)
from cognee.modules.retrie... | 119 | 3,978 |
cognee | cognee/eval_framework/beam/eval/run_sweep.py | .py | """Run BEAM answer/eval sweep against an already-ingested Cognee corpus."""
from __future__ import annotations
import argparse
import asyncio
import logging
import os
import time
from dataclasses import dataclass
from datetime import datetime, timezone
from pathlib import Path
import sys
from typing import Any, Optio... | 446 | 16,525 |
cognee | cognee/eval_framework/beam/eval/sweep.py | .py | from __future__ import annotations
import json
from copy import deepcopy
from datetime import datetime, timezone
from pathlib import Path
from typing import Any, Optional
from cognee.eval_framework.beam.eval.registry import ANSWERING_STRATEGIES
from cognee.eval_framework.eval_config import EvalConfig
from cognee.eval... | 231 | 8,252 |
cognee | cognee/eval_framework/beam/eval/metrics/kendall_tau.py | .py | """Kendall's tau-b metric for BEAM event_ordering questions.
Computes Kendall's tau-b rank correlation between the predicted and reference
event orderings, combined with F1 for event coverage.
Final score = tau_b_normalized × event_f1.
Only applies to event_ordering questions; returns None for all other types.
Refer... | 273 | 9,274 |
cognee | cognee/eval_framework/beam/eval/metrics/beam_rubric.py | .py | """BEAM rubric metric with 3-level scoring (0.0, 0.5, 1.0).
Uses the judge prompt from the official BEAM paper repository. Each rubric
criterion is scored independently by an LLM judge. Final score = mean across
all criteria for that question.
Reference: https://github.com/mohammadtavakoli78/BEAM/blob/main/src/evalua... | 241 | 9,483 |
cognee | cognee/eval_framework/corpus_builder/run_corpus_builder.py | .py | from cognee.shared.logging_utils import get_logger, ERROR
import json
from typing import List, Optional
from cognee.infrastructure.files.storage import get_file_storage
from cognee.eval_framework.corpus_builder.corpus_builder_executor import CorpusBuilderExecutor
from cognee.modules.data.models.questions_base import Q... | 78 | 3,049 |
cognee | cognee/eval_framework/corpus_builder/corpus_builder_executor.py | .py | import cognee
from cognee.shared.logging_utils import get_logger, ERROR
from typing import Optional, Tuple, List, Dict, Union, Any, Callable, Awaitable
from cognee.eval_framework.benchmark_adapters.benchmark_adapters import BenchmarkAdapter
from cognee.modules.chunking.TextChunker import TextChunker
from cognee.module... | 82 | 2,924 |
cognee | cognee/eval_framework/corpus_builder/task_getters/get_default_tasks_by_indices.py | .py | from typing import List
from cognee.api.v1.cognify.cognify import get_default_tasks
from cognee.modules.pipelines.tasks.task import Task
from cognee.modules.chunking.TextChunker import TextChunker
from cognee.tasks.graph import extract_graph_from_data
from cognee.tasks.storage import add_data_points
from cognee.shared.... | 61 | 2,325 |
cognee | cognee/eval_framework/corpus_builder/task_getters/TaskGetters.py | .py | from enum import Enum
from typing import Callable, Awaitable, List
from cognee.api.v1.cognify.cognify import get_default_tasks
from cognee.modules.pipelines.tasks.task import Task
from cognee.eval_framework.corpus_builder.task_getters.get_cascade_graph_tasks import (
get_cascade_graph_tasks,
)
from cognee.eval_fram... | 34 | 1,184 |
cognee | cognee/eval_framework/corpus_builder/task_getters/get_cascade_graph_tasks.py | .py | from typing import List
from pydantic import BaseModel
from cognee.modules.cognify.config import get_cognify_config
from cognee.modules.pipelines.tasks.task import Task
from cognee.modules.users.methods import get_default_user
from cognee.modules.users.models import User
from cognee.shared.data_models import Knowledge... | 50 | 1,830 |
cognee | cognee/eval_framework/evaluation/evaluator_adapters.py | .py | from enum import Enum
from importlib import import_module
from typing import Optional, Type
class EvaluatorAdapter(Enum):
"""Registry of evaluation engines.
The adapter class is resolved lazily through an import path instead of being
imported at module load time. This keeps optional dependencies (e.g.
... | 65 | 2,221 |
cognee | cognee/eval_framework/evaluation/direct_llm_eval_adapter.py | .py | from typing import Any, Dict, List
from pydantic import BaseModel
from cognee.eval_framework.evaluation.base_eval_adapter import BaseEvalAdapter
from cognee.eval_framework.eval_config import EvalConfig
from cognee.infrastructure.llm.prompts import render_prompt, read_query_prompt
from cognee.infrastructure.llm import ... | 60 | 2,167 |
cognee | cognee/eval_framework/evaluation/deep_eval_adapter.py | .py | from deepeval.metrics import GEval
from deepeval.test_case import LLMTestCase, LLMTestCaseParams
from cognee.eval_framework.eval_config import EvalConfig
from cognee.eval_framework.evaluation.base_eval_adapter import BaseEvalAdapter
from cognee.eval_framework.evaluation.metrics.exact_match import ExactMatchMetric
from ... | 106 | 4,532 |
cognee | cognee/eval_framework/evaluation/base_eval_adapter.py | .py | from abc import ABC, abstractmethod
from typing import Any, Dict, List
class BaseEvalAdapter(ABC):
@abstractmethod
async def evaluate_answers(
self, data: List[Dict[str, Any]], evaluator_metrics: List[str]
) -> List[Dict[str, Any]]:
pass
| 11 | 268 |
cognee | cognee/eval_framework/evaluation/run_evaluation_module.py | .py | from cognee.shared.logging_utils import get_logger
import json
from typing import List
from cognee.eval_framework.evaluation.evaluation_executor import EvaluationExecutor
from cognee.eval_framework.analysis.metrics_calculator import calculate_metrics_statistics
from cognee.infrastructure.files.storage import get_file_s... | 81 | 3,217 |
cognee | cognee/eval_framework/evaluation/evaluation_executor.py | .py | from typing import List, Dict, Any, Union
from cognee.eval_framework.evaluation.evaluator_adapters import EvaluatorAdapter
class EvaluationExecutor:
def __init__(
self,
evaluator_engine: Union[str, EvaluatorAdapter, Any] = "DeepEval",
evaluate_contexts: bool = False,
) -> None:
... | 29 | 1,208 |
cognee | cognee/eval_framework/evaluation/metrics/context_coverage.py | .py | from deepeval.metrics import SummarizationMetric
from deepeval.test_case import LLMTestCase
from deepeval.metrics.summarization.schema import ScoreType
from deepeval.metrics.indicator import metric_progress_indicator
from deepeval.utils import get_or_create_event_loop
class ContextCoverageMetric(SummarizationMetric):... | 51 | 2,028 |
cognee | cognee/eval_framework/evaluation/metrics/rubric.py | .py | """Rubric-based evaluation metric for BEAM benchmark.
Scores an LLM response against a list of rubric criteria using an LLM judge.
Each rubric item describes something the response "should contain" or "should state".
The score is the fraction of rubric items satisfied (0.0 to 1.0).
Unlike DeepEval's GEval, this metri... | 165 | 5,242 |
cognee | cognee/eval_framework/evaluation/metrics/f1.py | .py | from collections import Counter
from deepeval.test_case import LLMTestCase
import re
from typing import Optional, Any
class F1ScoreMetric:
def __init__(self) -> None:
self.score: Optional[float] = None
self.reason: Optional[str] = None
def measure(self, test_case: "LLMTestCase") -> float:
... | 47 | 1,652 |
cognee | cognee/eval_framework/evaluation/metrics/exact_match.py | .py | from deepeval.test_case import LLMTestCase
from typing import Optional
class ExactMatchMetric:
def __init__(self) -> None:
self.score: Optional[float] = None
self.reason: Optional[str] = None
def measure(self, test_case: "LLMTestCase") -> float:
actual = test_case.actual_output.strip(... | 16 | 628 |
cognee | cognee/eval_framework/answer_generation/run_question_answering_module.py | .py | from cognee.shared.logging_utils import get_logger
import json
from typing import List, Optional
from cognee.eval_framework.answer_generation.answer_generation_executor import (
AnswerGeneratorExecutor,
retriever_options,
)
from cognee.infrastructure.files.storage import get_file_storage
from cognee.infrastruct... | 77 | 2,995 |
cognee | cognee/eval_framework/answer_generation/beam_router.py | .py | """BEAM question-type router — routes probing questions to appropriate retrievers.
Each BEAM question type maps to a retriever + system prompt strategy.
The router classifies the question (using pre-labeled types from the dataset)
and delegates to the matching retrieval strategy.
"""
from typing import Any, Dict, Lis... | 190 | 8,034 |
cognee | cognee/eval_framework/answer_generation/answer_generation_executor.py | .py | from typing import List, Dict, Any
from cognee.modules.retrieval.completion_retriever import CompletionRetriever
from cognee.modules.retrieval.graph_completion_context_extension_retriever import (
GraphCompletionContextExtensionRetriever,
)
from cognee.modules.retrieval.graph_completion_cot_retriever import GraphCo... | 60 | 2,310 |
cognee | cognee/eval_framework/answer_generation/question_type_prompts.py | .py | from __future__ import annotations
from functools import lru_cache
from pathlib import Path
from typing import Mapping, Optional
QUESTION_TYPE_DEFAULT_KEY = "DEFAULT"
def get_question_type_prompt(
prompt_paths: Mapping[str, str],
question_type: str,
) -> Optional[str]:
prompt_path = prompt_paths.get(que... | 28 | 708 |
cognee | cognee/eval_framework/token_usage_analysis/cost_model.py | .py | """Cost comparison between full-context prompting and cognee persistent memory.
Pure arithmetic, no IO. The two querying strategies are modelled as objects that
each compute their own cumulative token cost over a number of queries. Every other
figure (parity, reduction milestones) is derived from those two objects.
""... | 115 | 4,137 |
cognee | cognee/eval_framework/token_usage_analysis/analyze.py | .py | """Estimate the token cost of cognee memory vs. full-context prompting.
Measure a few representative chunks of an input, then extrapolate the
full-context-vs-cognee cost comparison to the whole corpus. See README.md.
"""
from __future__ import annotations
import json
from pathlib import Path
from dotenv import load... | 47 | 1,263 |
cognee | cognee/eval_framework/token_usage_analysis/plot.py | .py | """Optional plotting: one cumulative-cost cross-over figure per llm_model.
Matplotlib is imported lazily so a JSON-only run needs no extra dependency. The
curves are rebuilt from the four numbers already in the report, so this stays a
pure consumer of the report with no cost-model knowledge.
"""
from __future__ impor... | 62 | 2,119 |
cognee | cognee/eval_framework/token_usage_analysis/measure.py | .py | """Measure the real token cost of ingesting a chunk with cognee.
This is the only module that calls the LLM. Each sampled chunk is run through
cognee's summary and graph-extraction calls, and the real prompt/completion
token usage is read off each response (so instruction and schema overhead are
included). Multiple ll... | 110 | 4,172 |
cognee | cognee/eval_framework/token_usage_analysis/cli.py | .py | """Command-line interface: argument definitions and default resolution.
Kept separate so analyze.py reads as pure orchestration.
"""
from __future__ import annotations
import argparse
from pathlib import Path
from corpus import DEFAULT_MAX_CHUNK_SIZE
DESCRIPTION = "Estimate the token cost of cognee memory vs. full... | 82 | 3,026 |
cognee | cognee/eval_framework/token_usage_analysis/corpus.py | .py | """Turn an input (file, directory, or raw string) into sampled chunks.
No LLM calls and no cost math here — just reading text and running cognee's
TextChunker so the sampled chunks match what ingestion would actually see.
"""
from __future__ import annotations
import asyncio
import importlib
import random
from pathl... | 74 | 2,771 |
cognee | cognee/eval_framework/token_usage_analysis/report.py | .py | """Per-llm_model orchestration: turn measurements into the JSON report.
Reads as: take this llm_model's chunk measurements, size the corpus, build the
two query-cost strategies, and read the reduction milestones off them.
"""
from __future__ import annotations
from cost_model import (
ChunkMeasurement,
Cogne... | 99 | 3,530 |
cognee | cognee/modules/session_distillation/models.py | .py | """Data models and tunables for session distillation.
Distillation turns a finished session's distillable context entries into standalone lesson
documents in the knowledge graph. The flow:
select -> batch (qa + candidates) -> curate per batch -> judge + write per lesson -> persist
Curator calls run in parallel b... | 99 | 3,270 |
cognee | cognee/modules/session_distillation/__init__.py | .py | from .distill import distill_session
from .models import (
CuratorBatchOutput,
DistillationResult,
ProposedLesson,
WrittenLesson,
)
__all__ = [
"distill_session",
"CuratorBatchOutput",
"DistillationResult",
"ProposedLesson",
"WrittenLesson",
]
| 16 | 281 |
cognee | cognee/modules/session_distillation/distill.py | .py | """Distill a finished session's learnings into the persistent knowledge graph.
Flow (curator calls parallel by batch; accept/write calls parallel by lesson):
1. LOAD session QA turns + distillable session-context entries.
2. CURATE pack the session timeline into batches; one curator LLM call per batch.
3. ACCEPT ... | 406 | 14,243 |
cognee | cognee/modules/settings/save_vector_db_config.py | .py | from typing import Union, Literal
from pydantic import BaseModel
from cognee.infrastructure.databases.vector import get_vectordb_config
class VectorDBConfig(BaseModel):
url: str
api_key: str
provider: Union[Literal["lancedb"], Literal["pgvector"]]
async def save_vector_db_config(vector_db_config: Vector... | 20 | 652 |
cognee | cognee/modules/settings/get_current_settings.py | .py | from typing import TypedDict
from cognee.infrastructure.llm import get_llm_config
from cognee.infrastructure.databases.graph import get_graph_config
from cognee.infrastructure.databases.vector import get_vectordb_config
from cognee.infrastructure.databases.relational.config import get_relational_config
class LLMConfi... | 61 | 1,642 |
cognee | cognee/modules/settings/get_settings.py | .py | from enum import Enum
from typing import Optional
from pydantic import BaseModel
from cognee.infrastructure.databases.vector import get_vectordb_config
from cognee.infrastructure.llm import get_llm_config
class ConfigChoice(BaseModel):
value: str
label: str
class ModelName(Enum):
openai = "openai"
o... | 192 | 5,820 |
cognee | cognee/modules/settings/save_llm_config.py | .py | import os
from pydantic import BaseModel
from cognee.infrastructure.llm import get_llm_config
class LLMConfig(BaseModel):
api_key: str
model: str
provider: str
async def save_llm_config(new_llm_config: LLMConfig):
llm_config = get_llm_config()
llm_config.llm_provider = new_llm_config.provider
... | 24 | 730 |
cognee | cognee/modules/provenance/storage.py | .py | """Async repository for the provenance ledger over the shared relational DB.
Hash-chain appends must be race-free under async SQLAlchemy on SQLite AND
Postgres, possibly multi-process. Three layers of defense:
1. A process-local per-event-loop ``asyncio.Lock`` (cheap serialization of the
common case).
2. A Postgre... | 302 | 12,722 |
cognee | cognee/modules/provenance/manager.py | .py | """Async ProvenanceManager over the audit-grade provenance ledger.
Port of semantica's ProvenanceManager onto cognee's shared relational DB
(SQLite + Postgres) via the async repository in ``storage.py``. The manager is
stateless (sessions per call), so acquisition is a trivial cached factory.
Tracking methods (``trac... | 708 | 29,036 |
cognee | cognee/modules/provenance/__init__.py | .py | """Audit-grade provenance ledger (append-only, tamper-evident).
This module is the fourth — and only audit-grade — provenance system in cognee.
It is deliberately distinct from the other three:
- ``cognee/infrastructure/databases/provenance/`` — graph source-refs used for
delete/rollback ownership (``source_ref:v1:... | 26 | 1,312 |
cognee | cognee/modules/provenance/integrity.py | .py | """Integrity primitives for the provenance ledger (pure, sync).
Checksum canonicalization is a deliberate break from semantica's byte format,
which had two defects: no-separator concatenation (field-boundary ambiguity)
and ``None`` hashing differently on the object vs dict path. Cognee ships no
semantica databases to ... | 126 | 4,570 |
cognee | cognee/modules/provenance/models/ProvenanceEntryRow.py | .py | from sqlalchemy import JSON, Boolean, Float, Index, Integer, String, Text, text
from sqlalchemy.orm import Mapped, mapped_column
from cognee.infrastructure.databases.relational import Base
class ProvenanceEntryRow(Base):
"""One row of the append-only provenance ledger.
PK is a plain string, not a UUID colum... | 95 | 5,091 |
cognee | cognee/modules/provenance/models/ProvenanceEntry.py | .py | from datetime import datetime, timezone
from typing import Any, Dict, List, Optional
from pydantic import BaseModel, ConfigDict, Field
from .ProvenanceEntryRow import ProvenanceEntryRow
def utc_now_iso() -> str:
"""Timezone-aware ISO-8601 timestamp (fixes semantica's naive-utcnow quirk)."""
return datetime.... | 107 | 3,794 |
cognee | cognee/modules/provenance/models/__init__.py | .py | """Provenance ledger models.
Importing this package registers ``ProvenanceEntryRow`` on the shared
``Base.metadata`` so both the startup ``create_all`` path and alembic's
``env.py`` see the ``provenance_entries`` table.
"""
from .ProvenanceEntry import ProvenanceEntry
from .ProvenanceEntryRow import ProvenanceEntryRo... | 12 | 375 |
cognee | cognee/modules/run_custom_pipeline/run_custom_pipeline.py | .py | from typing import Union, Optional, List, Type, Any
from uuid import UUID
from cognee.shared.logging_utils import get_logger
from cognee.modules.pipelines import run_pipeline
from cognee.modules.pipelines.tasks.task import Task
from cognee.modules.users.models import User
from cognee.modules.pipelines.layers.pipeline... | 87 | 4,632 |
cognee | cognee/modules/storage/utils/__init__.py | .py | import json
import copy
from uuid import UUID
from decimal import Decimal
from datetime import date, datetime
from pydantic_core import PydanticUndefined
from pydantic import create_model, ConfigDict, BaseModel, Field
from cognee.infrastructure.engine import DataPoint
class JSONEncoder(json.JSONEncoder):
def def... | 79 | 2,596 |
cognee | cognee/modules/chunking/Chunker.py | .py | class Chunker:
def __init__(self, document, get_text: callable, max_chunk_size: int):
self.chunk_index = 0
self.chunk_size = 0
self.token_count = 0
self.document = document
self.max_chunk_size = max_chunk_size
self.get_text = get_text
def read(self):
rai... | 13 | 343 |
cognee | cognee/modules/chunking/LangchainChunker.py | .py | from cognee.shared.logging_utils import get_logger
from os.path import basename
from uuid import NAMESPACE_OID, uuid5
from cognee.modules.chunking.Chunker import Chunker
from .models.DocumentChunk import DocumentChunk
from langchain_text_splitters import RecursiveCharacterTextSplitter
from cognee.infrastructure.databa... | 67 | 2,790 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.