datarobot-genai 0.2.39__py3-none-any.whl → 0.3.1__py3-none-any.whl

datarobot_genai/drmcp/tools/microsoft_graph/tools.py

@@ -16,6 +16,7 @@
 
 import logging
 from typing import Annotated
+from typing import Any
 from typing import Literal
 
 from fastmcp.exceptions import ToolError
@@ -124,10 +125,6 @@ async def microsoft_graph_search_content(
     - Documentation: https://learn.microsoft.com/en-us/graph/api/search-query
     - Search concepts: https://learn.microsoft.com/en-us/graph/search-concept-files
 
-    Permissions:
-    - Requires Sites.Read.All or Sites.Search.All permission
-    - include_hidden_content only works with delegated permissions
-    - region parameter is required for application permissions in multi-region environments
     """
     if not search_query:
         raise ToolError("Argument validation error: 'search_query' cannot be empty.")
@@ -171,12 +168,7 @@ async def microsoft_graph_search_content(
         }
         results.append(result_dict)
 
-    n = len(results)
     return ToolResult(
-        content=(
-            f"Successfully searched Microsoft Graph and retrieved {n} result(s) for "
-            f"'{search_query}' (from={from_offset}, size={size})."
-        ),
         structured_content={
             "query": search_query,
             "siteUrl": site_url,
@@ -184,7 +176,7 @@ async def microsoft_graph_search_content(
             "from": from_offset,
             "size": size,
             "results": results,
-            "count": n,
+            "count": len(results),
         },
     )
 
@@ -234,18 +226,12 @@ async def microsoft_graph_share_item(
             send_invitation=send_invitation,
         )
 
-    n = len(recipient_emails)
     return ToolResult(
-        content=(
-            f"Successfully shared file {file_id} "
-            f"from document library {document_library_id} "
-            f"with {n} recipients with '{role}' role."
-        ),
         structured_content={
             "fileId": file_id,
             "documentLibraryId": document_library_id,
             "recipientEmails": recipient_emails,
-            "n": n,
+            "n": len(recipient_emails),
             "role": role,
         },
     )
@@ -261,7 +247,8 @@ async def microsoft_graph_share_item(
         "create",
         "file",
         "write",
-    }
+    },
+    enabled=False,
 )
 async def microsoft_create_file(
     *,
@@ -318,7 +305,6 @@ async def microsoft_create_file(
         )
 
     return ToolResult(
-        content=f"File '{created_file.name}' created successfully.",
         structured_content={
             "file_name": created_file.name,
             "destination": "onedrive" if is_personal_onedrive else "sharepoint",
@@ -328,3 +314,133 @@ async def microsoft_create_file(
             "parentFolderId": created_file.parent_folder_id,
         },
     )
+
+
+@dr_mcp_tool(
+    tags={
+        "microsoft",
+        "graph api",
+        "sharepoint",
+        "onedrive",
+        "metadata",
+        "update",
+        "fields",
+        "compliance",
+    },
+    enabled=False,
+)
+async def microsoft_update_metadata(
+    *,
+    item_id: Annotated[str, "The ID of the file or list item to update."],
+    fields_to_update: Annotated[
+        dict[str, Any],
+        "Key-value pairs of metadata fields to modify. "
+        "For SharePoint list items: any custom column values. "
+        "For drive items: 'name' and/or 'description'.",
+    ],
+    site_id: Annotated[
+        str | None,
+        "The site ID (required for SharePoint list items, along with list_id).",
+    ] = None,
+    list_id: Annotated[
+        str | None,
+        "The list ID (required for SharePoint list items, along with site_id).",
+    ] = None,
+    document_library_id: Annotated[
+        str | None,
+        "The drive ID (required for OneDrive/drive item updates). "
+        "Cannot be used together with site_id and list_id.",
+    ] = None,
+) -> ToolResult | ToolError:
+    """
+    Update metadata on a SharePoint list item or OneDrive/SharePoint drive item.
+
+    **SharePoint List Items:** Provide site_id and list_id to update custom
+    column values on a list item. All custom columns can be updated.
+
+    **OneDrive/Drive Items:** Provide document_library_id to update drive item
+    properties. Only 'name' and 'description' fields can be updated.
+
+    **Context Requirements:**
+    - For SharePoint list items: Both site_id AND list_id are required
+    - For OneDrive/drive items: document_library_id is required
+    - Cannot specify both contexts simultaneously
+
+    **Examples:**
+    - SharePoint list item: Update a 'Status' column to 'Approved'
+    - Drive item: Rename a file or update its description
+
+    """
+    if not item_id or not item_id.strip():
+        raise ToolError("Error: item_id is required.")
+    if not fields_to_update:
+        raise ToolError("Error: fields_to_update is required and cannot be empty.")
+
+    # Validate context parameters
+    has_sharepoint_context = site_id is not None and list_id is not None
+    has_partial_sharepoint_context = (site_id is not None) != (list_id is not None)
+    has_drive_context = document_library_id is not None
+
+    if has_partial_sharepoint_context:
+        raise ToolError(
+            "Error: For SharePoint list items, both site_id and list_id must be provided."
+        )
+
+    if has_sharepoint_context and has_drive_context:
+        raise ToolError(
+            "Error: Cannot specify both SharePoint (site_id + list_id) and OneDrive "
+            "(document_library_id) context. Choose one."
+        )
+
+    if not has_sharepoint_context and not has_drive_context:
+        raise ToolError(
+            "Error: Must specify either SharePoint context (site_id + list_id) or "
+            "OneDrive context (document_library_id)."
+        )
+
+    access_token = await get_microsoft_graph_access_token()
+    if isinstance(access_token, ToolError):
+        raise access_token
+
+    async with MicrosoftGraphClient(access_token=access_token) as client:
+        result = await client.update_item_metadata(
+            item_id=item_id.strip(),
+            fields_to_update=fields_to_update,
+            site_id=site_id,
+            list_id=list_id,
+            drive_id=document_library_id,
+        )
+
+    context_type = "sharepoint_list_item" if has_sharepoint_context else "drive_item"
+    structured: dict[str, Any] = {
+        "item_id": item_id,
+        "context_type": context_type,
+        "fields_updated": list(fields_to_update.keys()),
+    }
+
+    # Add context-specific IDs for traceability
+    if has_sharepoint_context:
+        structured["site_id"] = site_id
+        structured["list_id"] = list_id
+    else:
+        structured["document_library_id"] = document_library_id
+
+    # Include relevant response data
+    if isinstance(result, dict):
+        # For drive items, include key properties if present
+        if has_drive_context:
+            if "id" in result:
+                structured["id"] = result["id"]
+            if "name" in result:
+                structured["name"] = result["name"]
+            if "webUrl" in result:
+                structured["webUrl"] = result["webUrl"]
+            if "description" in result:
+                structured["description"] = result.get("description")
+        # For list items, the response is the fields object itself
+        else:
+            structured["updated_fields"] = result
+
+    return ToolResult(
+        structured_content=structured,
+    )

datarobot_genai/drmcp/tools/perplexity/tools.py

@@ -0,0 +1,117 @@
+# Copyright 2026 DataRobot, Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Perplexity MCP tools."""
+
+import logging
+from typing import Annotated
+from typing import Literal
+
+from fastmcp.exceptions import ToolError
+from fastmcp.tools.tool import ToolResult
+
+from datarobot_genai.drmcp.core.mcp_instance import dr_mcp_tool
+from datarobot_genai.drmcp.tools.clients.perplexity import MAX_QUERIES
+from datarobot_genai.drmcp.tools.clients.perplexity import MAX_RESULTS
+from datarobot_genai.drmcp.tools.clients.perplexity import MAX_RESULTS_DEFAULT
+from datarobot_genai.drmcp.tools.clients.perplexity import MAX_SEARCH_DOMAIN_FILTER
+from datarobot_genai.drmcp.tools.clients.perplexity import MAX_TOKENS_PER_PAGE
+from datarobot_genai.drmcp.tools.clients.perplexity import MAX_TOKENS_PER_PAGE_DEFAULT
+from datarobot_genai.drmcp.tools.clients.perplexity import PerplexityClient
+from datarobot_genai.drmcp.tools.clients.perplexity import get_perplexity_access_token
+
+logger = logging.getLogger(__name__)
+
+
+@dr_mcp_tool(tags={"perplexity", "web", "search", "websearch"})
+async def perplexity_search(
+    *,
+    query: Annotated[
+        str,
+        list[str],
+        f"The search query string OR "
+        f"a list of up to {MAX_QUERIES} sub-queries for multi-query research.",
+    ],
+    search_domain_filter: Annotated[
+        list[str] | None,
+        f"Up to {MAX_SEARCH_DOMAIN_FILTER} domains/URLs "
+        f"to allowlist or denylist (prefix with '-').",
+    ] = None,
+    recency: Annotated[
+        Literal["day", "week", "month", "year"] | None, "Filter results by time period."
+    ] = None,
+    max_results: Annotated[
+        int, f"Number of ranked results to return (1-{MAX_RESULTS})."
+    ] = MAX_RESULTS_DEFAULT,
+    max_tokens_per_page: Annotated[
+        int,
+        f"Content extraction cap per page (1-{MAX_TOKENS_PER_PAGE}) "
+        f"(default {MAX_TOKENS_PER_PAGE_DEFAULT}).",
+    ] = MAX_TOKENS_PER_PAGE_DEFAULT,
+) -> ToolResult:
+    """Perplexity web search tool combining multi-query research and content extraction control."""
+    if not query:
+        raise ToolError("Argument validation error: query cannot be empty.")
+    if query and isinstance(query, str) and not query.strip():
+        raise ToolError("Argument validation error: query cannot be empty.")
+    if query and isinstance(query, list) and len(query) > MAX_QUERIES:
+        raise ToolError(
+            f"Argument validation error: query list cannot be bigger than {MAX_QUERIES}."
+        )
+    if query and isinstance(query, list) and not all(q.strip() for q in query):
+        raise ToolError("Argument validation error: query cannot contain empty str.")
+    if search_domain_filter and len(search_domain_filter) > MAX_SEARCH_DOMAIN_FILTER:
+        raise ToolError(
+            f"Argument validation error: "
+            f"maximum number of search domain filters is {MAX_SEARCH_DOMAIN_FILTER}."
+        )
+    if max_results <= 0:
+        raise ToolError("Argument validation error: max_results must be greater than 0.")
+    if max_results > MAX_RESULTS:
+        raise ToolError(
+            f"Argument validation error: "
+            f"max_results must be smaller than or equal to {MAX_RESULTS}."
+        )
+    if max_tokens_per_page <= 0:
+        raise ToolError("Argument validation error: max_tokens_per_page must be greater than 0.")
+    if max_tokens_per_page > MAX_TOKENS_PER_PAGE:
+        raise ToolError(
+            f"Argument validation error: "
+            f"max_tokens_per_page must be smaller than or equal to {MAX_TOKENS_PER_PAGE}."
+        )
+
+    access_token = await get_perplexity_access_token()
+    if isinstance(access_token, ToolError):
+        raise access_token
+
+    async with PerplexityClient(access_token=access_token) as perplexity_client:
+        results = await perplexity_client.search(
+            query=query,
+            search_domain_filter=search_domain_filter,
+            recency=recency,
+            max_results=max_results,
+            max_tokens_per_page=max_tokens_per_page,
+        )
+
+    return ToolResult(
+        structured_content={
+            "results": results,
+            "count": len(results),
+            "metadata": {
+                "queriesExecuted": len(query) if isinstance(query, list) else 1,
+                "filtersApplied": {"domains": search_domain_filter, "recency": recency},
+                "extractionLimit": max_tokens_per_page,
+            },
+        },
+    )

datarobot_genai/drmcp/tools/predictive/data.py

@@ -12,7 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import json
 import logging
 import os
 from typing import Annotated
@@ -60,7 +59,6 @@ async def upload_dataset_to_ai_catalog(
         raise ToolError("Failed to upload dataset.")
 
     return ToolResult(
-        content=f"Successfully uploaded dataset: {catalog_item.id}",
         structured_content={
             "dataset_id": catalog_item.id,
             "dataset_version_id": catalog_item.version_id,
@@ -78,21 +76,15 @@ async def list_ai_catalog_items() -> ToolResult:
     if not datasets:
         logger.info("No AI Catalog items found")
         return ToolResult(
-            content="No AI Catalog items found.",
             structured_content={"datasets": []},
         )
 
     datasets_dict = {ds.id: ds.name for ds in datasets}
-    datasets_count = len(datasets)
 
     return ToolResult(
-        content=(
-            f"Found {datasets_count} AI Catalog items, here are the details:\n"
-            f"{json.dumps(datasets_dict, indent=2)}"
-        ),
         structured_content={
             "datasets": datasets_dict,
-            "count": datasets_count,
+            "count": len(datasets),
         },
     )
 

datarobot_genai/drmcp/tools/predictive/deployment.py

@@ -12,7 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import json
 import logging
 from typing import Annotated
 
@@ -32,12 +31,10 @@ async def list_deployments() -> ToolResult:
     deployments = client.Deployment.list()
     if not deployments:
         return ToolResult(
-            content="No deployments found.",
             structured_content={"deployments": []},
         )
     deployments_dict = {d.id: d.label for d in deployments}
     return ToolResult(
-        content="\n".join(f"{d.id}: {d.label}" for d in deployments),
         structured_content={"deployments": deployments_dict},
     )
 
@@ -54,10 +51,6 @@ async def get_model_info_from_deployment(
     client = get_sdk_client()
     deployment = client.Deployment.get(deployment_id)
     return ToolResult(
-        content=(
-            f"Retrieved model info for deployment {deployment_id}, here are the details:\n"
-            f"{json.dumps(deployment.model, indent=2)}"
-        ),
         structured_content=deployment.model,
     )
 
@@ -87,7 +80,6 @@ async def deploy_model(
             default_prediction_server_id=prediction_servers[0].id,
         )
         return ToolResult(
-            content=f"Created deployment {deployment.id} with label {label}",
             structured_content={
                 "deployment_id": deployment.id,
                 "label": label,

datarobot_genai/drmcp/tools/predictive/deployment_info.py

@@ -95,7 +95,6 @@ async def get_deployment_info(
         }
 
     return ToolResult(
-        content=json.dumps(result, indent=2),
         structured_content=result,
     )
 
@@ -179,21 +178,6 @@ async def generate_prediction_data_template(
     # Create DataFrame
     df = pd.DataFrame(template_data)
 
-    # Add metadata comments
-    result = f"# Prediction Data Template for Deployment: {deployment_id}\n"
-    result += f"# Model Type: {features_info['model_type']}\n"
-    result += f"# Target: {features_info['target']} (Type: {features_info['target_type']})\n"
-
-    if "time_series_config" in features_info:
-        ts = features_info["time_series_config"]
-        result += f"# Time Series: datetime_column={ts['datetime_column']}, "
-        result += f"forecast_window=[{ts['forecast_window_start']}, {ts['forecast_window_end']}]\n"
-        if ts["series_id_columns"]:
-            result += f"# Multiseries ID Columns: {', '.join(ts['series_id_columns'])}\n"
-
-    result += f"# Total Features: {features_info['total_features']}\n"
-    result += df.to_csv(index=False)
-
     # Build structured content with template data and metadata
     structured_content = {
         "deployment_id": deployment_id,
@@ -208,7 +192,6 @@ async def generate_prediction_data_template(
         structured_content["time_series_config"] = features_info["time_series_config"]
 
     return ToolResult(
-        content=str(result),
         structured_content=structured_content,
     )
 
@@ -342,7 +325,6 @@ async def validate_prediction_data(
     }
 
     return ToolResult(
-        content=json.dumps(validation_report, indent=2),
         structured_content=validation_report,
     )
 
@@ -380,6 +362,5 @@ async def get_deployment_features(
         result["target_type"] = info["target_type"]
 
     return ToolResult(
-        content=json.dumps(result, indent=2),
         structured_content=result,
     )

datarobot_genai/drmcp/tools/predictive/model.py

@@ -93,33 +93,17 @@ async def get_best_model(
     best_model = leaderboard[0]
     logger.info(f"Found best model {best_model.id} for project {project_id}")
 
-    metric_info = ""
     metric_value = None
 
     if metric and best_model.metrics and metric in best_model.metrics:
         metric_value = best_model.metrics[metric].get("validation")
-        if metric_value is not None:
-            metric_info = f" with {metric}: {metric_value:.2f}"
 
     # Include full metrics in the response
     best_model_dict = model_to_dict(best_model)
     best_model_dict["metric"] = metric
     best_model_dict["metric_value"] = metric_value
 
-    # Format metrics for human-readable content
-    metrics_text = ""
-    if best_model.metrics:
-        metrics_list = []
-        for metric_name, metric_data in best_model.metrics.items():
-            if isinstance(metric_data, dict) and "validation" in metric_data:
-                val = metric_data["validation"]
-                if val is not None:
-                    metrics_list.append(f"{metric_name}: {val:.4f}")
-        if metrics_list:
-            metrics_text = "\nPerformance metrics:\n" + "\n".join(f"  - {m}" for m in metrics_list)
-
     return ToolResult(
-        content=f"Best model: {best_model.model_type}{metric_info}{metrics_text}",
         structured_content={
             "project_id": project_id,
             "best_model": best_model_dict,
@@ -148,7 +132,6 @@ async def score_dataset_with_model(
     job = model.score(dataset_url)
 
     return ToolResult(
-        content=f"Scoring job started: {job.id}",
         structured_content={
             "scoring_job_id": job.id,
             "project_id": project_id,
@@ -172,10 +155,6 @@ async def list_models(
     models = project.get_models()
 
     return ToolResult(
-        content=(
-            f"Found {len(models)} models in project {project_id}, here are the details:\n"
-            f"{json.dumps(models, indent=2, cls=ModelEncoder)}"
-        ),
         structured_content={
             "project_id": project_id,
             "models": [model_to_dict(model) for model in models],

datarobot_genai/drmcp/tools/predictive/predict_realtime.py

@@ -240,6 +240,9 @@ async def predict_realtime(
     else:
         raise ValueError("Either file_path or dataset must be provided.")
 
+    # Normalize column names: strip leading/trailing whitespace
+    df.columns = df.columns.str.strip()
+
     if series_id_column and series_id_column not in df.columns:
         raise ValueError(f"series_id_column '{series_id_column}' not found in input data.")
 

datarobot_genai/drmcp/tools/predictive/project.py

@@ -12,7 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import json
 import logging
 from typing import Annotated
 
@@ -33,11 +32,6 @@ async def list_projects() -> ToolResult:
     projects = {p.id: p.project_name for p in projects}
 
     return ToolResult(
-        content=(
-            json.dumps(projects, indent=2)
-            if projects
-            else json.dumps({"message": "No projects found."}, indent=2)
-        ),
         structured_content=projects,
     )
 
@@ -48,7 +42,7 @@ async def get_project_dataset_by_name(
     project_id: Annotated[str, "The ID of the DataRobot project."] | None = None,
     dataset_name: Annotated[str, "The name of the dataset to find (e.g., 'training', 'holdout')."]
     | None = None,
-) -> ToolError | ToolResult:
+) -> ToolResult:
     """Get a dataset ID by name for a given project.
 
     The dataset ID and the dataset type (source or prediction) as a string, or an error message.
@@ -70,21 +64,11 @@ async def get_project_dataset_by_name(
     for ds in all_datasets:
         if dataset_name.lower() in ds["dataset"].name.lower():
             return ToolResult(
-                content=(
-                    json.dumps(
-                        {
-                            "dataset_id": ds["dataset"].id,
-                            "dataset_type": ds["type"],
-                        },
-                        indent=2,
-                    )
-                ),
                 structured_content={
                     "dataset_id": ds["dataset"].id,
                     "dataset_type": ds["type"],
                 },
             )
-    return ToolResult(
-        content=f"Dataset with name containing '{dataset_name}' not found in project {project_id}.",
-        structured_content={},
+    raise ToolError(
+        f"Dataset with name containing '{dataset_name}' not found in project {project_id}."
     )

datarobot_genai/drmcp/tools/predictive/training.py

@@ -14,7 +14,6 @@
 
 """Tools for analyzing datasets and suggesting ML use cases."""
 
-import json
 import logging
 from dataclasses import asdict
 from dataclasses import dataclass
@@ -134,7 +133,6 @@ async def analyze_dataset(
     insights_dict = asdict(insights)
 
     return ToolResult(
-        content=json.dumps(insights_dict, indent=2),
         structured_content=insights_dict,
     )
 
@@ -164,7 +162,6 @@ async def suggest_use_cases(
     suggestions.sort(key=lambda x: x["confidence"], reverse=True)
 
     return ToolResult(
-        content=json.dumps(suggestions, indent=2),
         structured_content={"use_case_suggestions": suggestions},
     )
 
@@ -255,7 +252,6 @@ async def get_exploratory_insights(
                 )
 
     return ToolResult(
-        content=json.dumps(eda_insights, indent=2),
         structured_content=eda_insights,
     )
 
@@ -540,22 +536,11 @@ async def start_autopilot(
         }
 
         return ToolResult(
-            content=json.dumps(result, indent=2),
             structured_content=result,
         )
 
     except Exception as e:
-        raise ToolError(
-            content=json.dumps(
-                {
-                    "error": f"Failed to start Autopilot: {str(e)}",
-                    "project_id": project.id if project else None,
-                    "target": target,
-                    "mode": mode,
-                },
-                indent=2,
-            )
-        )
+        raise ToolError(f"Failed to start Autopilot: {str(e)}")
 
 
 @dr_mcp_tool(tags={"prediction", "training", "read", "model", "evaluation"})
@@ -611,7 +596,6 @@ async def get_model_roc_curve(
         }
 
         return ToolResult(
-            content=json.dumps({"data": roc_data}, indent=2),
             structured_content={"data": roc_data},
         )
     except Exception as e:
@@ -638,7 +622,6 @@ async def get_model_feature_impact(
     feature_impact = model.get_or_request_feature_impact()
 
     return ToolResult(
-        content=json.dumps({"data": feature_impact}, indent=2),
         structured_content={"data": feature_impact},
     )
 
@@ -684,6 +667,5 @@ async def get_model_lift_chart(
     }
 
     return ToolResult(
-        content=json.dumps({"data": lift_chart_data}, indent=2),
         structured_content={"data": lift_chart_data},
     )

datarobot_genai/drmcp/tools/tavily/__init__.py

@@ -0,0 +1,13 @@
+# Copyright 2025 DataRobot, Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.