morphik 1.1.2__tar.gz → 1.2.0__tar.gz

{morphik-1.1.2 → morphik-1.2.0}/.gitignore

@@ -58,5 +58,5 @@ multi_vector_embeddings_*.json
 # Rust build artifacts
 morphik_rust/target/
 morphik_rust/Cargo.lock
-testkb_morphik/*
-DOCS_BY_FACILITIES/*
+DOCS_BY_FACILITIES/
+eval_freshworks/

{morphik-1.1.2 → morphik-1.2.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: morphik
-Version: 1.1.2
+Version: 1.2.0
 Summary: Morphik Python Client
 Author-email: Morphik <founders@morphik.ai>
 Requires-Python: >=3.8
@@ -101,7 +101,7 @@ chunks = folder.retrieve_chunks(query="design notes", folder_depth=-1)
 docs = db.list_documents(folder_name="/projects/alpha", folder_depth=-1)
 ```
 
-`Folder.full_path` is exposed on folder objects, and `Document.folder_path` / `Graph.folder_path` mirror server responses for tracing scope.
+`Folder.full_path` is exposed on folder objects, and `Document.folder_path` mirrors server responses for tracing scope.
 
 ### Asynchronous Usage
 
@@ -145,7 +145,6 @@ asyncio.run(main())
 - Semantic search and retrieval
 - Retrieval-augmented generation (RAG)
 - Morphik On-the-Fly document querying with optional ingestion follow-up
-- Knowledge graph creation and querying
 - Multi-user and multi-folder scoping
 - Metadata filtering
 - Document management

{morphik-1.1.2 → morphik-1.2.0}/README.md

@@ -88,7 +88,7 @@ chunks = folder.retrieve_chunks(query="design notes", folder_depth=-1)
 docs = db.list_documents(folder_name="/projects/alpha", folder_depth=-1)
 ```
 
-`Folder.full_path` is exposed on folder objects, and `Document.folder_path` / `Graph.folder_path` mirror server responses for tracing scope.
+`Folder.full_path` is exposed on folder objects, and `Document.folder_path` mirrors server responses for tracing scope.
 
 ### Asynchronous Usage
 
@@ -132,7 +132,6 @@ asyncio.run(main())
 - Semantic search and retrieval
 - Retrieval-augmented generation (RAG)
 - Morphik On-the-Fly document querying with optional ingestion follow-up
-- Knowledge graph creation and querying
 - Multi-user and multi-folder scoping
 - Metadata filtering
 - Document management

{morphik-1.1.2 → morphik-1.2.0}/morphik/__init__.py

@@ -14,4 +14,4 @@ __all__ = [
     "DocumentQueryResponse",
 ]
 
-__version__ = "1.1.2"
+__version__ = "1.2.0"

{morphik-1.1.2 → morphik-1.2.0}/morphik/_internal.py

@@ -1,7 +1,6 @@
 import base64
 import io
 import json
-import warnings
 from datetime import date, datetime
 from decimal import Decimal
 from io import BytesIO
@@ -21,15 +20,9 @@ from .models import (
     Document,
     DocumentQueryResponse,
     DocumentResult,
-    Graph,
-    GraphPromptOverrides,
     IngestTextRequest,
     QueryPromptOverrides,
 )
-from .rules import Rule
-
-# Type alias for rules
-RuleOrDict = Union[Rule, Dict[str, Any]]
 
 
 class FinalChunkResult(BaseModel):
@@ -81,21 +74,6 @@ class _MorphikClientLogic:
         # Basic token validation
         jwt.decode(self._auth_token, options={"verify_signature": False})
 
-    def _convert_rule(self, rule: RuleOrDict) -> Dict[str, Any]:
-        """Convert a rule to a dictionary format"""
-        if hasattr(rule, "to_dict"):
-            return rule.to_dict()
-        return rule
-
-    def _warn_legacy_rules(self, rules: Optional[List[RuleOrDict]], context: str) -> None:
-        """Emit a deprecation warning when legacy rules are supplied."""
-        if rules:
-            warnings.warn(
-                f"'rules' support has been removed; payload supplied to {context} is ignored.",
-                DeprecationWarning,
-                stacklevel=3,
-            )
-
     def _get_url(self, endpoint: str) -> str:
         """Get the full URL for an API endpoint"""
         return f"{self._base_url}/{endpoint.lstrip('/')}"
@@ -112,13 +90,11 @@ class _MorphikClientLogic:
         content: str,
         filename: Optional[str],
         metadata: Optional[Dict[str, Any]],
-        rules: Optional[List[RuleOrDict]],
         use_colpali: bool,
         folder_name: Optional[str],
         end_user_id: Optional[str],
     ) -> Dict[str, Any]:
         """Prepare request for ingest_text endpoint"""
-        self._warn_legacy_rules(rules, "ingest/text")
         serialized_metadata, metadata_types_map = self._serialize_metadata_map(metadata)
         payload = {
             "content": content,
@@ -185,7 +161,6 @@ class _MorphikClientLogic:
     def _prepare_ingest_file_form_data(
         self,
         metadata: Optional[Dict[str, Any]],
-        rules: Optional[List[RuleOrDict]],
         folder_name: Optional[str],
         end_user_id: Optional[str],
         use_colpali: Optional[bool] = None,
@@ -196,7 +171,6 @@ class _MorphikClientLogic:
         never relies on query-string values.  *use_colpali* is therefore always
         embedded here when provided.
         """
-        self._warn_legacy_rules(rules, "ingest/file")
         serialized_metadata, metadata_types_map = self._serialize_metadata_map(metadata)
         form_data = {
             "metadata": json.dumps(serialized_metadata),
@@ -219,15 +193,12 @@ class _MorphikClientLogic:
     def _prepare_ingest_files_form_data(
         self,
         metadata: Optional[Union[Dict[str, Any], List[Dict[str, Any]]]],
-        rules: Optional[List[RuleOrDict]],
         use_colpali: bool,
         parallel: bool,
         folder_name: Optional[str],
         end_user_id: Optional[str],
     ) -> Dict[str, Any]:
         """Prepare form data for ingest_files endpoint"""
-        self._warn_legacy_rules(rules, "ingest/files")
-
         serialized_metadata, metadata_types_payload = self._serialize_metadata_collection(metadata)
 
         data = {
@@ -290,9 +261,6 @@ class _MorphikClientLogic:
         max_tokens: Optional[int],
         temperature: Optional[float],
         use_colpali: bool,
-        graph_name: Optional[str],
-        hop_depth: int,
-        include_paths: bool,
         prompt_overrides: Optional[Union[QueryPromptOverrides, Dict[str, Any]]],
         folder_name: Optional[Union[str, List[str]]],
         folder_depth: Optional[int],
@@ -317,9 +285,6 @@ class _MorphikClientLogic:
             "temperature": temperature,
             "use_colpali": use_colpali,
             "use_reranking": use_reranking,  # Add to payload
-            "graph_name": graph_name,
-            "hop_depth": hop_depth,
-            "include_paths": include_paths,
             "prompt_overrides": prompt_overrides,
         }
         if folder_name:
@@ -506,69 +471,15 @@ class _MorphikClientLogic:
             request["output_format"] = output_format
         return request
 
-    def _prepare_create_graph_request(
-        self,
-        name: str,
-        filters: Optional[Dict[str, Any]],
-        documents: Optional[List[str]],
-        prompt_overrides: Optional[Union[GraphPromptOverrides, Dict[str, Any]]],
-        folder_name: Optional[Union[str, List[str]]],
-        end_user_id: Optional[str],
-    ) -> Dict[str, Any]:
-        """Prepare request for create_graph endpoint"""
-        # Convert prompt_overrides to dict if it's a model
-        if prompt_overrides and isinstance(prompt_overrides, GraphPromptOverrides):
-            prompt_overrides = prompt_overrides.model_dump(exclude_none=True)
-
-        request = {
-            "name": name,
-            "filters": filters,
-            "documents": documents,
-            "prompt_overrides": prompt_overrides,
-        }
-        if folder_name:
-            request["folder_name"] = folder_name
-        if end_user_id:
-            request["end_user_id"] = end_user_id
-        return request
-
-    def _prepare_update_graph_request(
-        self,
-        name: str,
-        additional_filters: Optional[Dict[str, Any]],
-        additional_documents: Optional[List[str]],
-        prompt_overrides: Optional[Union[GraphPromptOverrides, Dict[str, Any]]],
-        folder_name: Optional[Union[str, List[str]]],
-        end_user_id: Optional[str],
-    ) -> Dict[str, Any]:
-        """Prepare request for update_graph endpoint"""
-        # Convert prompt_overrides to dict if it's a model
-        if prompt_overrides and isinstance(prompt_overrides, GraphPromptOverrides):
-            prompt_overrides = prompt_overrides.model_dump(exclude_none=True)
-
-        request = {
-            "additional_filters": additional_filters,
-            "additional_documents": additional_documents,
-            "prompt_overrides": prompt_overrides,
-        }
-        if folder_name:
-            request["folder_name"] = folder_name
-        if end_user_id:
-            request["end_user_id"] = end_user_id
-        return request
-
     def _prepare_update_document_with_text_request(
         self,
         document_id: str,
         content: str,
         filename: Optional[str],
         metadata: Optional[Dict[str, Any]],
-        rules: Optional[List],
         use_colpali: Optional[bool],
     ) -> Tuple[Dict[str, Any], Dict[str, Any]]:
         """Prepare request for update_document_with_text endpoint"""
-        self._warn_legacy_rules(rules, "documents/update_text")
-
         serialized_metadata, metadata_types_map = self._serialize_metadata_map(metadata)
         request = IngestTextRequest(
             content=content,
@@ -721,14 +632,6 @@ class _MorphikClientLogic:
 
         return final_chunks
 
-    def _parse_graph_response(self, response_json: Dict[str, Any]) -> Graph:
-        """Parse graph response"""
-        return Graph(**response_json)
-
-    def _parse_graph_list_response(self, response_json: List[Dict[str, Any]]) -> List[Graph]:
-        """Parse graph list response"""
-        return [Graph(**graph) for graph in response_json]
-
     def _parse_document_query_response(self, response_json: Dict[str, Any]) -> DocumentQueryResponse:
         """Parse document query response."""
         payload = dict(response_json)

{morphik-1.1.2 → morphik-1.2.0}/morphik/_scoped_ops.py

@@ -5,7 +5,7 @@ from typing import Any, Callable, Dict, List, Optional, Type, TypeVar, Union
 
 from pydantic import BaseModel
 
-from ._internal import FinalChunkResult, RuleOrDict
+from ._internal import FinalChunkResult
 
 T = TypeVar("T")
 
@@ -39,7 +39,6 @@ class _ScopedOperationsMixin:
         content: str,
         filename: Optional[str],
         metadata: Optional[Dict[str, Any]],
-        rules: Optional[List[RuleOrDict]],
         use_colpali: bool,
         folder_name: Optional[str],
         end_user_id: Optional[str],
@@ -48,7 +47,6 @@ class _ScopedOperationsMixin:
             content,
             filename,
             metadata,
-            rules,
             use_colpali,
             folder_name,
             end_user_id,
@@ -67,7 +65,6 @@ class _ScopedOperationsMixin:
         file: Union[str, bytes, BytesIO, IOBase, Path],
         filename: Optional[str],
         metadata: Optional[Dict[str, Any]],
-        rules: Optional[List[RuleOrDict]],
         use_colpali: bool,
         folder_name: Optional[str],
         end_user_id: Optional[str],
@@ -80,7 +77,6 @@ class _ScopedOperationsMixin:
 
         form_data = self._logic._prepare_ingest_file_form_data(
             metadata,
-            rules,
             folder_name,
             end_user_id,
             use_colpali,
@@ -100,7 +96,6 @@ class _ScopedOperationsMixin:
         *,
         files: List[Union[str, bytes, BytesIO, IOBase, Path]],
         metadata: Optional[Union[Dict[str, Any], List[Dict[str, Any]]]],
-        rules: Optional[List[RuleOrDict]],
         use_colpali: bool,
         parallel: bool,
         folder_name: Optional[str],
@@ -115,7 +110,6 @@ class _ScopedOperationsMixin:
 
         form_data = self._logic._prepare_ingest_files_form_data(
             metadata,
-            rules,
             use_colpali,
             parallel,
             folder_name,
@@ -222,9 +216,6 @@ class _ScopedOperationsMixin:
         max_tokens: Optional[int],
         temperature: Optional[float],
         use_colpali: bool,
-        graph_name: Optional[str],
-        hop_depth: int,
-        include_paths: bool,
         prompt_overrides: Optional[Dict[str, Any]],
         folder_name: Optional[Union[str, List[str]]],
         folder_depth: Optional[int],
@@ -243,9 +234,6 @@ class _ScopedOperationsMixin:
             max_tokens,
             temperature,
             use_colpali,
-            graph_name,
-            hop_depth,
-            include_paths,
             prompt_overrides,
             folder_name,
             folder_depth,

morphik-1.2.0/morphik/_shared.py

@@ -0,0 +1,182 @@
+from __future__ import annotations
+
+import json
+from pathlib import Path
+from typing import Any, Dict, Iterable, List, Optional, Union
+
+from pydantic import BaseModel
+
+MAX_LIMIT = 500
+MIN_LOG_HOURS = 0.1
+MAX_LOG_HOURS = 168.0
+
+
+def merge_folders(
+    base: Optional[Union[str, List[str]]],
+    additional: Optional[List[str]],
+) -> Optional[Union[str, List[str]]]:
+    if not additional:
+        return base
+    if base:
+        if isinstance(base, list):
+            return base + additional
+        return [base] + additional
+    return additional
+
+
+def collect_directory_files(directory: Union[str, Path], recursive: bool, pattern: str) -> List[Path]:
+    dir_path = Path(directory)
+    if not dir_path.is_dir():
+        raise ValueError(f"Directory not found: {dir_path}")
+
+    files = list(dir_path.rglob(pattern) if recursive else dir_path.glob(pattern))
+    return [f for f in files if f.is_file()]
+
+
+def normalize_limit_offset(limit: int, offset: int) -> Dict[str, int]:
+    return {
+        "limit": max(1, min(limit, MAX_LIMIT)),
+        "offset": max(0, offset),
+    }
+
+
+def normalize_filter_param(value: Optional[Union[str, Dict[str, Any], List[Any]]]) -> Optional[str]:
+    if value is None:
+        return None
+    if isinstance(value, str):
+        return value
+    return json.dumps(value)
+
+
+def build_list_apps_params(
+    *,
+    org_id: Optional[str],
+    user_id: Optional[str],
+    app_id_filter: Optional[Union[str, Dict[str, Any], List[Any]]],
+    app_name_filter: Optional[Union[str, Dict[str, Any], List[Any]]],
+    limit: int,
+    offset: int,
+) -> Dict[str, Any]:
+    params: Dict[str, Any] = normalize_limit_offset(limit, offset)
+    if org_id:
+        params["org_id"] = org_id
+    if user_id:
+        params["user_id"] = user_id
+    if app_id_filter is not None:
+        params["app_id_filter"] = normalize_filter_param(app_id_filter)
+    if app_name_filter is not None:
+        params["app_name_filter"] = normalize_filter_param(app_name_filter)
+    return params
+
+
+def build_rename_app_params(*, new_name: str, app_id: Optional[str], app_name: Optional[str]) -> Dict[str, Any]:
+    if not app_id and not app_name:
+        raise ValueError("app_id or app_name is required to rename an app")
+    params: Dict[str, Any] = {"new_name": new_name}
+    if app_id:
+        params["app_id"] = app_id
+    if app_name:
+        params["app_name"] = app_name
+    return params
+
+
+def build_rotate_app_params(
+    *, app_id: Optional[str], app_name: Optional[str], expiry_days: Optional[int]
+) -> Dict[str, Any]:
+    if not app_id and not app_name:
+        raise ValueError("app_id or app_name is required to rotate an app token")
+    params: Dict[str, Any] = {}
+    if app_id:
+        params["app_id"] = app_id
+    if app_name:
+        params["app_name"] = app_name
+    if expiry_days is not None:
+        params["expiry_days"] = expiry_days
+    return params
+
+
+def build_create_app_payload(
+    *,
+    name: str,
+    app_id: Optional[str],
+    user_id: Optional[str],
+    expiry_days: Optional[int],
+    org_id: Optional[str],
+    created_by_user_id: Optional[str],
+) -> Dict[str, Any]:
+    payload: Dict[str, Any] = {"name": name}
+    if app_id is not None:
+        payload["app_id"] = app_id
+    if user_id is not None:
+        payload["user_id"] = user_id
+    if expiry_days is not None:
+        payload["expiry_days"] = expiry_days
+    if org_id is not None:
+        payload["org_id"] = org_id
+    if created_by_user_id is not None:
+        payload["created_by_user_id"] = created_by_user_id
+    return payload
+
+
+def build_requeue_payload(
+    *,
+    jobs: Optional[Iterable[Union[BaseModel, Dict[str, Any]]]],
+    include_all: bool,
+    statuses: Optional[List[str]],
+    limit: Optional[int],
+) -> Dict[str, Any]:
+    jobs_list = list(jobs) if jobs is not None else None
+    if not include_all and not jobs_list:
+        raise ValueError("jobs or include_all must be provided for requeue")
+    payload: Dict[str, Any] = {}
+    if jobs_list:
+        payload["jobs"] = [job.model_dump() if isinstance(job, BaseModel) else job for job in jobs_list]
+    if include_all:
+        payload["include_all"] = True
+    if statuses is not None:
+        payload["statuses"] = statuses
+    if limit is not None:
+        payload["limit"] = limit
+    return payload
+
+
+def build_logs_params(
+    *, limit: int, hours: float, op_type: Optional[str], status: Optional[str]
+) -> Dict[str, Any]:
+    params: Dict[str, Any] = {
+        "limit": max(1, min(limit, MAX_LIMIT)),
+        "hours": max(MIN_LOG_HOURS, min(hours, MAX_LOG_HOURS)),
+    }
+    if op_type is not None:
+        params["op_type"] = op_type
+    if status is not None:
+        params["status"] = status
+    return params
+
+
+def build_document_by_filename_params(
+    *, folder_name: Optional[Union[str, List[str]]], folder_depth: Optional[int], end_user_id: Optional[str]
+) -> Dict[str, Any]:
+    params: Dict[str, Any] = {}
+    if folder_name is not None:
+        params["folder_name"] = folder_name
+    if folder_depth is not None:
+        params["folder_depth"] = folder_depth
+    if end_user_id is not None:
+        params["end_user_id"] = end_user_id
+    return params
+
+
+def normalize_additional_folders(
+    additional_folders: Optional[List[str]],
+    folder_name: Optional[Union[str, List[str]]],
+) -> Optional[List[str]]:
+    if folder_name is None:
+        return additional_folders
+    if isinstance(folder_name, str):
+        folder_list = [folder_name]
+    else:
+        folder_list = list(folder_name)
+    if additional_folders:
+        return list(additional_folders) + folder_list
+    return folder_list