cognee 0.4.1__py3-none-any.whl → 0.5.0.dev0__py3-none-any.whl

cognee/__init__.py

@@ -19,6 +19,7 @@ from .api.v1.add import add
 from .api.v1.delete import delete
 from .api.v1.cognify import cognify
 from .modules.memify import memify
+from .modules.run_custom_pipeline import run_custom_pipeline
 from .api.v1.update import update
 from .api.v1.config.config import config
 from .api.v1.datasets.datasets import datasets

cognee/api/client.py

@@ -23,6 +23,7 @@ from cognee.api.v1.settings.routers import get_settings_router
 from cognee.api.v1.datasets.routers import get_datasets_router
 from cognee.api.v1.cognify.routers import get_code_pipeline_router, get_cognify_router
 from cognee.api.v1.search.routers import get_search_router
+from cognee.api.v1.ontologies.routers.get_ontology_router import get_ontology_router
 from cognee.api.v1.memify.routers import get_memify_router
 from cognee.api.v1.add.routers import get_add_router
 from cognee.api.v1.delete.routers import get_delete_router
@@ -39,6 +40,8 @@ from cognee.api.v1.users.routers import (
 )
 from cognee.modules.users.methods.get_authenticated_user import REQUIRE_AUTHENTICATION
 
+# Ensure application logging is configured for container stdout/stderr
+setup_logging()
 logger = get_logger()
 
 if os.getenv("ENV", "prod") == "prod":
@@ -74,6 +77,9 @@ async def lifespan(app: FastAPI):
 
     await get_default_user()
 
+    # Emit a clear startup message for docker logs
+    logger.info("Backend server has started")
+
     yield
 
 
@@ -258,6 +264,8 @@ app.include_router(
 
 app.include_router(get_datasets_router(), prefix="/api/v1/datasets", tags=["datasets"])
 
+app.include_router(get_ontology_router(), prefix="/api/v1/ontologies", tags=["ontologies"])
+
 app.include_router(get_settings_router(), prefix="/api/v1/settings", tags=["settings"])
 
 app.include_router(get_visualize_router(), prefix="/api/v1/visualize", tags=["visualize"])

cognee/api/v1/add/routers/get_add_router.py

@@ -82,7 +82,9 @@ def get_add_router() -> APIRouter:
                 datasetName,
                 user=user,
                 dataset_id=datasetId,
-                node_set=node_set if node_set else None,
+                node_set=node_set
+                if node_set != [""]
+                else None,  # Transform default node_set endpoint value to None
             )
 
             if isinstance(add_run, PipelineRunErrored):

cognee/api/v1/cognify/routers/get_cognify_router.py

@@ -41,6 +41,9 @@ class CognifyPayloadDTO(InDTO):
     custom_prompt: Optional[str] = Field(
         default="", description="Custom prompt for entity extraction and graph generation"
     )
+    ontology_key: Optional[List[str]] = Field(
+        default=None, description="Reference to one or more previously uploaded ontologies"
+    )
 
 
 def get_cognify_router() -> APIRouter:
@@ -68,6 +71,7 @@ def get_cognify_router() -> APIRouter:
         - **dataset_ids** (Optional[List[UUID]]): List of existing dataset UUIDs to process. UUIDs allow processing of datasets not owned by the user (if permitted).
         - **run_in_background** (Optional[bool]): Whether to execute processing asynchronously. Defaults to False (blocking).
         - **custom_prompt** (Optional[str]): Custom prompt for entity extraction and graph generation. If provided, this prompt will be used instead of the default prompts for knowledge graph extraction.
+        - **ontology_key** (Optional[List[str]]): Reference to one or more previously uploaded ontology files to use for knowledge graph construction.
 
         ## Response
         - **Blocking execution**: Complete pipeline run information with entity counts, processing duration, and success/failure status
@@ -82,7 +86,8 @@ def get_cognify_router() -> APIRouter:
         {
             "datasets": ["research_papers", "documentation"],
             "run_in_background": false,
-            "custom_prompt": "Extract entities focusing on technical concepts and their relationships. Identify key technologies, methodologies, and their interconnections."
+            "custom_prompt": "Extract entities focusing on technical concepts and their relationships. Identify key technologies, methodologies, and their interconnections.",
+            "ontology_key": ["medical_ontology_v1"]
         }
         ```
 
@@ -108,13 +113,35 @@ def get_cognify_router() -> APIRouter:
             )
 
         from cognee.api.v1.cognify import cognify as cognee_cognify
+        from cognee.api.v1.ontologies.ontologies import OntologyService
 
         try:
             datasets = payload.dataset_ids if payload.dataset_ids else payload.datasets
+            config_to_use = None
+
+            if payload.ontology_key:
+                ontology_service = OntologyService()
+                ontology_contents = ontology_service.get_ontology_contents(
+                    payload.ontology_key, user
+                )
+
+                from cognee.modules.ontology.ontology_config import Config
+                from cognee.modules.ontology.rdf_xml.RDFLibOntologyResolver import (
+                    RDFLibOntologyResolver,
+                )
+                from io import StringIO
+
+                ontology_streams = [StringIO(content) for content in ontology_contents]
+                config_to_use: Config = {
+                    "ontology_config": {
+                        "ontology_resolver": RDFLibOntologyResolver(ontology_file=ontology_streams)
+                    }
+                }
 
             cognify_run = await cognee_cognify(
                 datasets,
                 user,
+                config=config_to_use,
                 run_in_background=payload.run_in_background,
                 custom_prompt=payload.custom_prompt,
             )

cognee/api/v1/ontologies/__init__.py

@@ -0,0 +1,4 @@
+from .ontologies import OntologyService
+from .routers.get_ontology_router import get_ontology_router
+
+__all__ = ["OntologyService", "get_ontology_router"]

cognee/api/v1/ontologies/ontologies.py

@@ -0,0 +1,183 @@
+import os
+import json
+import tempfile
+from pathlib import Path
+from datetime import datetime, timezone
+from typing import Optional, List
+from dataclasses import dataclass
+
+
+@dataclass
+class OntologyMetadata:
+    ontology_key: str
+    filename: str
+    size_bytes: int
+    uploaded_at: str
+    description: Optional[str] = None
+
+
+class OntologyService:
+    def __init__(self):
+        pass
+
+    @property
+    def base_dir(self) -> Path:
+        return Path(tempfile.gettempdir()) / "ontologies"
+
+    def _get_user_dir(self, user_id: str) -> Path:
+        user_dir = self.base_dir / str(user_id)
+        user_dir.mkdir(parents=True, exist_ok=True)
+        return user_dir
+
+    def _get_metadata_path(self, user_dir: Path) -> Path:
+        return user_dir / "metadata.json"
+
+    def _load_metadata(self, user_dir: Path) -> dict:
+        metadata_path = self._get_metadata_path(user_dir)
+        if metadata_path.exists():
+            with open(metadata_path, "r") as f:
+                return json.load(f)
+        return {}
+
+    def _save_metadata(self, user_dir: Path, metadata: dict):
+        metadata_path = self._get_metadata_path(user_dir)
+        with open(metadata_path, "w") as f:
+            json.dump(metadata, f, indent=2)
+
+    async def upload_ontology(
+        self, ontology_key: str, file, user, description: Optional[str] = None
+    ) -> OntologyMetadata:
+        if not file.filename.lower().endswith(".owl"):
+            raise ValueError("File must be in .owl format")
+
+        user_dir = self._get_user_dir(str(user.id))
+        metadata = self._load_metadata(user_dir)
+
+        if ontology_key in metadata:
+            raise ValueError(f"Ontology key '{ontology_key}' already exists")
+
+        content = await file.read()
+        if len(content) > 10 * 1024 * 1024:
+            raise ValueError("File size exceeds 10MB limit")
+
+        file_path = user_dir / f"{ontology_key}.owl"
+        with open(file_path, "wb") as f:
+            f.write(content)
+
+        ontology_metadata = {
+            "filename": file.filename,
+            "size_bytes": len(content),
+            "uploaded_at": datetime.now(timezone.utc).isoformat(),
+            "description": description,
+        }
+        metadata[ontology_key] = ontology_metadata
+        self._save_metadata(user_dir, metadata)
+
+        return OntologyMetadata(
+            ontology_key=ontology_key,
+            filename=file.filename,
+            size_bytes=len(content),
+            uploaded_at=ontology_metadata["uploaded_at"],
+            description=description,
+        )
+
+    async def upload_ontologies(
+        self, ontology_key: List[str], files: List, user, descriptions: Optional[List[str]] = None
+    ) -> List[OntologyMetadata]:
+        """
+        Upload ontology files with their respective keys.
+
+        Args:
+            ontology_key: List of unique keys for each ontology
+            files: List of UploadFile objects (same length as keys)
+            user: Authenticated user
+            descriptions: Optional list of descriptions for each file
+
+        Returns:
+            List of OntologyMetadata objects for uploaded files
+
+        Raises:
+            ValueError: If keys duplicate, file format invalid, or array lengths don't match
+        """
+        if len(ontology_key) != len(files):
+            raise ValueError("Number of keys must match number of files")
+
+        if len(set(ontology_key)) != len(ontology_key):
+            raise ValueError("Duplicate ontology keys not allowed")
+
+        if descriptions and len(descriptions) != len(files):
+            raise ValueError("Number of descriptions must match number of files")
+
+        results = []
+        user_dir = self._get_user_dir(str(user.id))
+        metadata = self._load_metadata(user_dir)
+
+        for i, (key, file) in enumerate(zip(ontology_key, files)):
+            if key in metadata:
+                raise ValueError(f"Ontology key '{key}' already exists")
+
+            if not file.filename.lower().endswith(".owl"):
+                raise ValueError(f"File '{file.filename}' must be in .owl format")
+
+            content = await file.read()
+            if len(content) > 10 * 1024 * 1024:
+                raise ValueError(f"File '{file.filename}' exceeds 10MB limit")
+
+            file_path = user_dir / f"{key}.owl"
+            with open(file_path, "wb") as f:
+                f.write(content)
+
+            ontology_metadata = {
+                "filename": file.filename,
+                "size_bytes": len(content),
+                "uploaded_at": datetime.now(timezone.utc).isoformat(),
+                "description": descriptions[i] if descriptions else None,
+            }
+            metadata[key] = ontology_metadata
+
+            results.append(
+                OntologyMetadata(
+                    ontology_key=key,
+                    filename=file.filename,
+                    size_bytes=len(content),
+                    uploaded_at=ontology_metadata["uploaded_at"],
+                    description=descriptions[i] if descriptions else None,
+                )
+            )
+
+        self._save_metadata(user_dir, metadata)
+        return results
+
+    def get_ontology_contents(self, ontology_key: List[str], user) -> List[str]:
+        """
+        Retrieve ontology content for one or more keys.
+
+        Args:
+            ontology_key: List of ontology keys to retrieve (can contain single item)
+            user: Authenticated user
+
+        Returns:
+            List of ontology content strings
+
+        Raises:
+            ValueError: If any ontology key not found
+        """
+        user_dir = self._get_user_dir(str(user.id))
+        metadata = self._load_metadata(user_dir)
+
+        contents = []
+        for key in ontology_key:
+            if key not in metadata:
+                raise ValueError(f"Ontology key '{key}' not found")
+
+            file_path = user_dir / f"{key}.owl"
+            if not file_path.exists():
+                raise ValueError(f"Ontology file for key '{key}' not found")
+
+            with open(file_path, "r", encoding="utf-8") as f:
+                contents.append(f.read())
+        return contents
+
+    def list_ontologies(self, user) -> dict:
+        user_dir = self._get_user_dir(str(user.id))
+        return self._load_metadata(user_dir)

cognee/api/v1/ontologies/routers/get_ontology_router.py

@@ -0,0 +1,107 @@
+from fastapi import APIRouter, File, Form, UploadFile, Depends, HTTPException
+from fastapi.responses import JSONResponse
+from typing import Optional, List
+
+from cognee.modules.users.models import User
+from cognee.modules.users.methods import get_authenticated_user
+from cognee.shared.utils import send_telemetry
+from cognee import __version__ as cognee_version
+from ..ontologies import OntologyService
+
+
+def get_ontology_router() -> APIRouter:
+    router = APIRouter()
+    ontology_service = OntologyService()
+
+    @router.post("", response_model=dict)
+    async def upload_ontology(
+        ontology_key: str = Form(...),
+        ontology_file: List[UploadFile] = File(...),
+        descriptions: Optional[str] = Form(None),
+        user: User = Depends(get_authenticated_user),
+    ):
+        """
+        Upload ontology files with their respective keys for later use in cognify operations.
+
+        Supports both single and multiple file uploads:
+        - Single file: ontology_key=["key"], ontology_file=[file]
+        - Multiple files: ontology_key=["key1", "key2"], ontology_file=[file1, file2]
+
+        ## Request Parameters
+        - **ontology_key** (str): JSON array string of user-defined identifiers for the ontologies
+        - **ontology_file** (List[UploadFile]): OWL format ontology files
+        - **descriptions** (Optional[str]): JSON array string of optional descriptions
+
+        ## Response
+        Returns metadata about uploaded ontologies including keys, filenames, sizes, and upload timestamps.
+
+        ## Error Codes
+        - **400 Bad Request**: Invalid file format, duplicate keys, array length mismatches, file size exceeded
+        - **500 Internal Server Error**: File system or processing errors
+        """
+        send_telemetry(
+            "Ontology Upload API Endpoint Invoked",
+            user.id,
+            additional_properties={
+                "endpoint": "POST /api/v1/ontologies",
+                "cognee_version": cognee_version,
+            },
+        )
+
+        try:
+            import json
+
+            ontology_keys = json.loads(ontology_key)
+            description_list = json.loads(descriptions) if descriptions else None
+
+            if not isinstance(ontology_keys, list):
+                raise ValueError("ontology_key must be a JSON array")
+
+            results = await ontology_service.upload_ontologies(
+                ontology_keys, ontology_file, user, description_list
+            )
+
+            return {
+                "uploaded_ontologies": [
+                    {
+                        "ontology_key": result.ontology_key,
+                        "filename": result.filename,
+                        "size_bytes": result.size_bytes,
+                        "uploaded_at": result.uploaded_at,
+                        "description": result.description,
+                    }
+                    for result in results
+                ]
+            }
+        except (json.JSONDecodeError, ValueError) as e:
+            return JSONResponse(status_code=400, content={"error": str(e)})
+        except Exception as e:
+            return JSONResponse(status_code=500, content={"error": str(e)})
+
+    @router.get("", response_model=dict)
+    async def list_ontologies(user: User = Depends(get_authenticated_user)):
+        """
+        List all uploaded ontologies for the authenticated user.
+
+        ## Response
+        Returns a dictionary mapping ontology keys to their metadata including filename, size, and upload timestamp.
+
+        ## Error Codes
+        - **500 Internal Server Error**: File system or processing errors
+        """
+        send_telemetry(
+            "Ontology List API Endpoint Invoked",
+            user.id,
+            additional_properties={
+                "endpoint": "GET /api/v1/ontologies",
+                "cognee_version": cognee_version,
+            },
+        )
+
+        try:
+            metadata = ontology_service.list_ontologies(user)
+            return metadata
+        except Exception as e:
+            return JSONResponse(status_code=500, content={"error": str(e)})
+
+    return router

cognee/api/v1/permissions/routers/get_permissions_router.py

@@ -1,15 +1,20 @@
 from uuid import UUID
-from typing import List
+from typing import List, Union
 
 from fastapi import APIRouter, Depends
 from fastapi.responses import JSONResponse
 
 from cognee.modules.users.models import User
+from cognee.api.DTO import InDTO
 from cognee.modules.users.methods import get_authenticated_user
 from cognee.shared.utils import send_telemetry
 from cognee import __version__ as cognee_version
 
 
+class SelectTenantDTO(InDTO):
+    tenant_id: UUID | None = None
+
+
 def get_permissions_router() -> APIRouter:
     permissions_router = APIRouter()
 
@@ -226,4 +231,39 @@ def get_permissions_router() -> APIRouter:
             status_code=200, content={"message": "Tenant created.", "tenant_id": str(tenant_id)}
         )
 
+    @permissions_router.post("/tenants/select")
+    async def select_tenant(payload: SelectTenantDTO, user: User = Depends(get_authenticated_user)):
+        """
+        Select current tenant.
+
+        This endpoint selects a tenant with the specified UUID. Tenants are used
+        to organize users and resources in multi-tenant environments, providing
+        isolation and access control between different groups or organizations.
+
+        Sending a null/None value as tenant_id selects his default single user tenant
+
+        ## Request Parameters
+        - **tenant_id** (Union[UUID, None]): UUID of the tenant to select, If null/None is provided use the default single user tenant
+
+        ## Response
+        Returns a success message along with selected tenant id.
+        """
+        send_telemetry(
+            "Permissions API Endpoint Invoked",
+            user.id,
+            additional_properties={
+                "endpoint": f"POST /v1/permissions/tenants/{str(payload.tenant_id)}",
+                "tenant_id": str(payload.tenant_id),
+            },
+        )
+
+        from cognee.modules.users.tenants.methods import select_tenant as select_tenant_method
+
+        await select_tenant_method(user_id=user.id, tenant_id=payload.tenant_id)
+
+        return JSONResponse(
+            status_code=200,
+            content={"message": "Tenant selected.", "tenant_id": str(payload.tenant_id)},
+        )
+
     return permissions_router

cognee/cli/commands/cognify_command.py

@@ -22,7 +22,7 @@ relationships, and creates semantic connections for enhanced search and reasonin
 
 Processing Pipeline:
 1. **Document Classification**: Identifies document types and structures
-2. **Permission Validation**: Ensures user has processing rights  
+2. **Permission Validation**: Ensures user has processing rights
 3. **Text Chunking**: Breaks content into semantically meaningful segments
 4. **Entity Extraction**: Identifies key concepts, people, places, organizations
 5. **Relationship Detection**: Discovers connections between entities
@@ -97,6 +97,13 @@ After successful cognify processing, use `cognee search` to query the knowledge
                             chunker_class = LangchainChunker
                         except ImportError:
                             fmt.warning("LangchainChunker not available, using TextChunker")
+                    elif args.chunker == "CsvChunker":
+                        try:
+                            from cognee.modules.chunking.CsvChunker import CsvChunker
+
+                            chunker_class = CsvChunker
+                        except ImportError:
+                            fmt.warning("CsvChunker not available, using TextChunker")
 
                     result = await cognee.cognify(
                         datasets=datasets,

cognee/cli/config.py

@@ -26,7 +26,7 @@ SEARCH_TYPE_CHOICES = [
 ]
 
 # Chunker choices
-CHUNKER_CHOICES = ["TextChunker", "LangchainChunker"]
+CHUNKER_CHOICES = ["TextChunker", "LangchainChunker", "CsvChunker"]
 
 # Output format choices
 OUTPUT_FORMAT_CHOICES = ["json", "pretty", "simple"]

cognee/context_global_variables.py

@@ -4,6 +4,8 @@ from typing import Union
 from uuid import UUID
 
 from cognee.base_config import get_base_config
+from cognee.infrastructure.databases.vector.config import get_vectordb_context_config
+from cognee.infrastructure.databases.graph.config import get_graph_context_config
 from cognee.infrastructure.databases.utils import get_or_create_dataset_database
 from cognee.infrastructure.files.storage.config import file_storage_config
 from cognee.modules.users.methods import get_user
@@ -14,11 +16,40 @@ vector_db_config = ContextVar("vector_db_config", default=None)
 graph_db_config = ContextVar("graph_db_config", default=None)
 session_user = ContextVar("session_user", default=None)
 
+VECTOR_DBS_WITH_MULTI_USER_SUPPORT = ["lancedb", "falkor"]
+GRAPH_DBS_WITH_MULTI_USER_SUPPORT = ["kuzu", "falkor"]
+
 
 async def set_session_user_context_variable(user):
     session_user.set(user)
 
 
+def multi_user_support_possible():
+    graph_db_config = get_graph_context_config()
+    vector_db_config = get_vectordb_context_config()
+    return (
+        graph_db_config["graph_database_provider"] in GRAPH_DBS_WITH_MULTI_USER_SUPPORT
+        and vector_db_config["vector_db_provider"] in VECTOR_DBS_WITH_MULTI_USER_SUPPORT
+    )
+
+
+def backend_access_control_enabled():
+    backend_access_control = os.environ.get("ENABLE_BACKEND_ACCESS_CONTROL", None)
+    if backend_access_control is None:
+        # If backend access control is not defined in environment variables,
+        # enable it by default if graph and vector DBs can support it, otherwise disable it
+        return multi_user_support_possible()
+    elif backend_access_control.lower() == "true":
+        # If enabled, ensure that the current graph and vector DBs can support it
+        multi_user_support = multi_user_support_possible()
+        if not multi_user_support:
+            raise EnvironmentError(
+                "ENABLE_BACKEND_ACCESS_CONTROL is set to true but the current graph and/or vector databases do not support multi-user access control. Please use supported databases or disable backend access control."
+            )
+        return True
+    return False
+
+
 async def set_database_global_context_variables(dataset: Union[str, UUID], user_id: UUID):
     """
     If backend access control is enabled this function will ensure all datasets have their own databases,
@@ -38,9 +69,7 @@ async def set_database_global_context_variables(dataset: Union[str, UUID], user_
 
     """
 
-    base_config = get_base_config()
-
-    if not os.getenv("ENABLE_BACKEND_ACCESS_CONTROL", "false").lower() == "true":
+    if not backend_access_control_enabled():
         return
 
     user = await get_user(user_id)
@@ -48,6 +77,7 @@ async def set_database_global_context_variables(dataset: Union[str, UUID], user_
     # To ensure permissions are enforced properly all datasets will have their own databases
     dataset_database = await get_or_create_dataset_database(dataset, user)
 
+    base_config = get_base_config()
     data_root_directory = os.path.join(
         base_config.data_root_directory, str(user.tenant_id or user.id)
     )
@@ -57,15 +87,17 @@ async def set_database_global_context_variables(dataset: Union[str, UUID], user_
 
     # Set vector and graph database configuration based on dataset database information
     vector_config = {
-        "vector_db_url": os.path.join(
-            databases_directory_path, dataset_database.vector_database_name
-        ),
-        "vector_db_key": "",
-        "vector_db_provider": "lancedb",
+        "vector_db_provider": dataset_database.vector_database_provider,
+        "vector_db_url": dataset_database.vector_database_url,
+        "vector_db_key": dataset_database.vector_database_key,
+        "vector_db_name": dataset_database.vector_database_name,
     }
 
     graph_config = {
-        "graph_database_provider": "kuzu",
+        "graph_database_provider": dataset_database.graph_database_provider,
+        "graph_database_url": dataset_database.graph_database_url,
+        "graph_database_name": dataset_database.graph_database_name,
+        "graph_database_key": dataset_database.graph_database_key,
         "graph_file_path": os.path.join(
             databases_directory_path, dataset_database.graph_database_name
         ),

cognee/infrastructure/databases/cache/config.py

@@ -1,6 +1,6 @@
 from pydantic_settings import BaseSettings, SettingsConfigDict
 from functools import lru_cache
-from typing import Optional
+from typing import Optional, Literal
 
 
 class CacheConfig(BaseSettings):
@@ -15,6 +15,7 @@ class CacheConfig(BaseSettings):
     - agentic_lock_timeout: Maximum time (in seconds) to wait for the lock release.
     """
 
+    cache_backend: Literal["redis", "fs"] = "fs"
     caching: bool = False
     shared_kuzu_lock: bool = False
     cache_host: str = "localhost"
@@ -28,6 +29,7 @@ class CacheConfig(BaseSettings):
 
     def to_dict(self) -> dict:
         return {
+            "cache_backend": self.cache_backend,
             "caching": self.caching,
             "shared_kuzu_lock": self.shared_kuzu_lock,
             "cache_host": self.cache_host,