OpenBudget

Dependency "fastmcp@0.3.0" has known CVEs:

Dependency "requests@2.31.0" has known CVEs:

Pattern "(?:fetch|axios|requests?\.get|http\.get).*(?:return|respond|result|body|text|data)(?!.*(?:sanitize|escape|encode|strip|validate|parse|extract))" matched in source_code: "Fetching data" (at position 5409)

Dependency "starlette@0.35.0" has known CVEs:

Pattern "\brequests\.(?:get|post|put|delete|patch|head|options|request)\s*\(\s*(?:req|request|input|param|params|args|url|uri|href|link|target|destination|endpoint|host|address|resource|src|source)" matched in source_code: "requests.get(url" (at position 5459)

Pattern "(?:readFile|read_file|query|select|getSecret|getCredential|getPassword).*(?:webhook|http|fetch|axios|post|send|email|slack|discord)" matched in source_code: "query: Post" (at position 10904)

Pattern "(?:delay|sleep|timeout|interval)\s*[:=]\s*(?:[^;]*(?:secret|token|password|credential|key|env))" matched in source_code: "timeout=30) response.raise_for_status() return response.json() except Exception as e: log.error(f"Error fetching dataset info for {dataset}: {e}") return {"error": str(e)} @mcp.tool(annotations=ANNOTATIONS) def DatasetFullTextSearch(dataset: str, q: str) -> dict: """ Perform full-text search on a dataset to locate relevant items and identifiers. **Purpose**: Use this tool to find specific textual identifiers (IDs, codes, names) needed for precise queries. Do NOT use this tool for searching time periods or dates. **Usage Instructions**: 1. ALWAYS call DatasetInfo first to understand the dataset structure 2. Use free-text search to find entities by name, title, or description 3. Extract relevant identifiers (like entity_id, code, budget_code) from the results 4. Use these identifiers in your DatasetDBQuery calls for precise filtering 5. NEVER present search results directly to the user as your final answer **When to use**: - To find the entity_id of an organization mentioned by name - To find budget item codes by searching for keywords - To locate contracts by supplier name or description - When you need specific identifiers but only have descriptive text **Important**: - Search results are for YOUR use to find identifiers, not for presenting to users - Always follow up with DatasetDBQuery using the identifiers you found - If unsure which identifier to use in a query, search first - AVOID calling more than 4 tools in parallel to prevent memory overflow Args: dataset: ID of the dataset to search. Available datasets: - budget_items_data: Budget book (ספר התקציב) - support_programs_data: Support programs (תוכניות תמיכה) - supports_transactions_data: Support payments (תשלומי תמיכות) - contracts_data: Procurement contracts (התקשרויות רכש) - entities_data: Organizations and entities (גופים וארגונים) - income_items_data: State revenues (הכנסות המדינה) - budgetary_change_requests_data: Budget change requests (בקשות שינוי תקציב) - budgetary_change_transactions_data: Budget change transactions (שינויי תקציב) q: Free-text search query (organization name, keyword, description, etc.) Returns: Search results with matching items and their identifiers """ try: url = f"{BUDGETKEY_API_BASE}/api/tables/{dataset}/search" params = {"q": q} log.info(f"Searching dataset {dataset}: {url}?q={q}") response = requests.get(url, params=params, timeout=30) response.raise_for_status() return response.json() except Exception as e: log.error(f"Error searching dataset {dataset} with query '{q}': {e}") return {"error": str(e)} @mcp.tool(annotations=ANNOTATIONS) def DatasetDBQuery(dataset: str, query: str, page_size: int = 50) -> dict: """ Execute PostgreSQL-compatible SQL queries to obtain comprehensive, precise information from datasets. **CRITICAL Prerequisites**: 1. MUST call DatasetInfo first to understand the dataset schema 2. Use exact column names from the schema (case-sensitive) 3. If you need identifiers, call DatasetFullTextSearch first to find them **Usage Instructions**: - Use only identifiers found through DatasetFullTextSearch - NEVER guess identifiers - Filter by relevant time periods (year, date fields) - Use aggregate functions (SUM, COUNT, AVG) to summarize data when appropriate - ALWAYS include the `item_url` field in SELECT to provide direct links to data - Construct queries based on the exact schema from DatasetInfo **Query Best Practices**: - Format results in tables when possible - Use ORDER BY to sort results meaningfully - Apply WHERE clauses for time periods and specific filters - Use JOINs when querying related datasets - Include descriptive fields (title, name, description) along with values **Handling Results**: - Check for warnings in the response - if present, fix the query and re-run - Extract the `download_url` field from results - Offer users download links formatted as markdown: [Download data](download_url) - Use `item_url` to create clickable links: [Item Name](item_url) - Present data professionally in tables or structured format **Important**: - NEVER present results based on a query that returned warnings - Always verify your SQL syntax is PostgreSQL-compatible - Use aggregate functions for summary data - Filter by time periods appropriately Args: dataset: ID of the dataset to query. Available datasets: - budget_items_data: Budget book (ספר התקציב) - support_programs_data: Support programs (תוכניות תמיכה) - supports_transactions_data: Support payments (תשלומי תמיכות) - contracts_data: Procurement contracts (התקשרויות רכש) - entities_data: Organizations and entities (גופים וארגונים) - income_items_data: State revenues (הכנסות המדינה) - budgetary_change_requests_data: Budget change requests (בקשות שינוי תקציב) - budgetary_change_transactions_data: Budget change transactions (שינויי תקציב) query: PostgreSQL-compatible SQL query to execute. Example: "SELECT year, code, title, net_allocated, net_executed, item_url FROM budget_items_data WHERE year = 2025 AND title LIKE '%חינוך%' ORDER BY net_allocated DESC LIMIT 10" page_size: Number of rows to return (default: 50) Returns: Query results including: - rows: Array of result rows - download_url: Link to download full dataset (offer this to users as markdown link) - warnings: Any warnings about the query (must fix and re-run if present) """ try: url = f"{BUDGETKEY" (at position 5477)