Fix filter parameter validation issue

- Fix schema inconsistency: change 'filters' (object) to 'filter' (string)
- Update tool schema to use proper OData string format
- Update README documentation to match corrected schema
- Add demo script showing old vs new format

Resolves filter validation error:
Before: {"filters": {"data_source": {"eq": "GAO"}}}
After: {"filter": "data_source eq 'GAO'"}

This fixes the "Filter validation failed" error users were encountering.

Author: dividor

README.md

@@ -126,13 +126,13 @@ For Docker:
 The server provides two main tools:
 
 ### 1. PIA Search
-Comprehensive search with OData filtering and faceting. The `filters` parameter uses standard [OData query syntax](https://docs.oasis-open.org/odata/odata/v4.01/odata-v4.01-part2-url-conventions.html).
+Comprehensive search with OData filtering and faceting. The `filter` parameter uses standard [OData query syntax](https://docs.oasis-open.org/odata/odata/v4.01/odata-v4.01-part2-url-conventions.html).
 
 **Tool Name:** `pia_search`
 
 **Parameters:**
 - `query` (required): Search query text
-- `filters` (optional): OData filter expression
+- `filter` (optional): OData filter expression
 - `page` (optional): Page number (default: 1)
 - `page_size` (optional): Results per page (default: 10)
 - `search_mode` (optional): Search mode (default: "content")
@@ -181,7 +181,7 @@ Discover available field names and values for filtering.
 - Find possible field values (e.g., "OIG", "GAO", "audit_report")
 - Understand data types for each field (string, date, number)
 
-This information helps you construct proper `filters` for the `pia_search` tool.
+This information helps you construct proper `filter` expressions for the `pia_search` tool.
 
 ## 🔍 Filter Discovery Workflow
 
@@ -207,13 +207,13 @@ Use the `pia_search` tool with discovered fields to create precise OData filters
 **Basic Example:**
 ```
 Query: "Medicare fraud"
-Filters: "data_source in ('OIG', 'CMS') and published_date ge '2023-01-01' and document_type eq 'audit_report'"
+Filter: "data_source in ('OIG', 'CMS') and published_date ge '2023-01-01' and document_type eq 'audit_report'"
 ```
 
 **Complex Example:**
 ```
 Query: "healthcare violations"
-Filters: "(data_source eq 'OIG' or data_source eq 'CMS') and (severity eq 'High' or amount gt 1000000) and published_date ge '2023-01-01'"
+Filter: "(data_source eq 'OIG' or data_source eq 'CMS') and (severity eq 'High' or amount gt 1000000) and published_date ge '2023-01-01'"
 ```
 
 ## 📝 AI Instruction Prompts

filter_fix_demo.py

@@ -0,0 +1,47 @@
+#!/usr/bin/env python3
+"""
+Demo script showing the correct filter format after the fix.
+
+BEFORE (caused error):
+{
+  "limit": 15,
+  "query": "ambulance fraud Medicare Medicaid emergency medical transportation",
+  "filters": {
+    "data_source": {
+      "eq": "GAO"
+    }
+  }
+}
+
+AFTER (correct format):
+{
+  "limit": 15,
+  "query": "ambulance fraud Medicare Medicaid emergency medical transportation",
+  "filter": "data_source eq 'GAO'"
+}
+"""
+
+import json
+
+# The problematic query that was causing the error
+old_format = {
+    "limit": 15,
+    "query": "ambulance fraud Medicare Medicaid emergency medical transportation",
+    "filters": {"data_source": {"eq": "GAO"}},
+}
+
+# The correct format after our fix
+new_format = {
+    "limit": 15,
+    "query": "ambulance fraud Medicare Medicaid emergency medical transportation",
+    "filter": "data_source eq 'GAO'",
+}
+
+print("❌ OLD FORMAT (caused Filter validation failed error):")
+print(json.dumps(old_format, indent=2))
+print("\n✅ NEW FORMAT (correct OData syntax):")
+print(json.dumps(new_format, indent=2))
+print("\n🔧 What changed:")
+print("1. Parameter name: 'filters' → 'filter'")
+print("2. Format: object with nested structure → OData string")
+print('3. Value: {"data_source": {"eq": "GAO"}} → "data_source eq \'GAO\'"')

src/pia_mcp_server/tools/search_tools.py

@@ -13,14 +13,14 @@
 # Tool definitions based on the API response
 pia_search_tool = types.Tool(
     name="pia_search",
-    description="Search the Program Integrity Alliance (PIA) database for documents and recommendations. Returns comprehensive results with full citation information and clickable links for proper attribution. Each result includes corresponding citations with data source attribution (GAO, OIG, etc.). Supports structured filtering using index field names with operators like eq, ne, in, not_in, gte, lte, etc.",
+    description="Search the Program Integrity Alliance (PIA) database for documents and recommendations. Returns comprehensive results with full citation information and clickable links for proper attribution. Each result includes corresponding citations with data source attribution (GAO, OIG, etc.). Supports OData filter expressions using operators like eq, ne, gt, ge, lt, le, and, or, etc.",
     inputSchema={
         "type": "object",
         "properties": {
             "query": {"type": "string", "description": "Search query text"},
-            "filters": {
-                "type": "object",
-                "description": "Optional filters for narrowing results",
+            "filter": {
+                "type": "string",
+                "description": "Optional OData filter expression for narrowing results (e.g., \"data_source eq 'GAO'\")",
             },
             "page": {
                 "type": "integer",