pinescript-mcp 0.7.2__tar.gz → 0.7.4__tar.gz

{pinescript_mcp-0.7.2 → pinescript_mcp-0.7.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pinescript-mcp
-Version: 0.7.2
+Version: 0.7.4
 Summary: MCP server providing Pine Script v6 documentation for AI assistants
 Project-URL: Homepage, https://github.com/paulieb89/pinescript-mcp
 Project-URL: Documentation, https://github.com/paulieb89/pinescript-mcp#readme
@@ -121,7 +121,7 @@ Documentation is bundled in the package — each version contains a frozen snaps
   "mcpServers": {
     "pinescript-docs": {
       "command": "uvx",
-      "args": ["pinescript-mcp==0.7.0"]
+      "args": ["pinescript-mcp==0.7.4"]
     }
   }
 }

{pinescript_mcp-0.7.2 → pinescript_mcp-0.7.4}/README.md

@@ -91,7 +91,7 @@ Documentation is bundled in the package — each version contains a frozen snaps
   "mcpServers": {
     "pinescript-docs": {
       "command": "uvx",
-      "args": ["pinescript-mcp==0.7.0"]
+      "args": ["pinescript-mcp==0.7.4"]
     }
   }
 }

{pinescript_mcp-0.7.2 → pinescript_mcp-0.7.4}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "pinescript-mcp"
-version = "0.7.2"
+version = "0.7.4"
 description = "MCP server providing Pine Script v6 documentation for AI assistants"
 readme = "README.md"
 license = "MIT"

{pinescript_mcp-0.7.2 → pinescript_mcp-0.7.4}/src/pinescript_mcp/__init__.py

@@ -4,4 +4,4 @@ Provides tools to list, search, and read Pine Script v6 documentation
 for use with AI assistants like Claude.
 """
 
-__version__ = "0.7.1"
+__version__ = "0.7.4"

{pinescript_mcp-0.7.2 → pinescript_mcp-0.7.4}/src/pinescript_mcp/server.py

@@ -11,6 +11,7 @@ from pathlib import Path
 from typing import Literal
 
 from fastmcp import FastMCP, Context
+from fastmcp.exceptions import ToolError
 from fastmcp.server.context import _current_transport
 from fastmcp.server.middleware.logging import StructuredLoggingMiddleware
 from fastmcp.server.middleware.rate_limiting import RateLimitingMiddleware
@@ -325,6 +326,15 @@ TOPIC_MAP = {
     "migration": "reference/migration_v5_to_v6.md",
 }
 
+# Known invalid/renamed functions → specific replacement hints for validate_function()
+KNOWN_REPLACEMENTS: dict[str, str] = {
+    "ta.adx": "ta.adx() does NOT exist. Use ta.dmi(diLen, adxSmoothing) → returns [diPlus, diMinus, adx] as a tuple.",
+    "ta.sum": "ta.sum() does NOT exist. Use math.sum(source, length) instead.",
+    "security": "security() was renamed in v5. Use request.security() instead.",
+    "study": "study() was renamed in v5. Use indicator() instead.",
+    "input": "input() works but prefer typed variants: input.int(), input.float(), input.string(), input.bool(), etc.",
+}
+
 
 def _find_section(content: str, header: str, include_children: bool = True) -> tuple[str, int, int]:
     """Find a section in markdown content by header text.
@@ -470,7 +480,7 @@ async def list_sections(path: str):
             return "\n".join(headers)
         except ValueError as e:
             log["error"] = str(e)
-            return f"Error: {e}"
+            raise ToolError(str(e))
 
 
 @mcp.tool(
@@ -503,7 +513,7 @@ async def get_doc(path: str, limit: int = 0, offset: int = 0):
 
             if limit > 0:
                 if offset >= total:
-                    return f"Error: offset {offset} exceeds file size ({total} chars). Use offset < {total}."
+                    raise ToolError(f"offset {offset} exceeds file size ({total} chars). Use offset < {total}.")
                 end = min(offset + limit, total)
                 content = content[offset:end]
                 has_more = end < total
@@ -515,7 +525,7 @@ async def get_doc(path: str, limit: int = 0, offset: int = 0):
                 return content
         except ValueError as e:
             log["error"] = str(e)
-            return f"Error: {e}"
+            raise ToolError(str(e))
 
 
 @mcp.tool(
@@ -543,7 +553,7 @@ async def get_section(path: str, header: str, include_children: bool = True):
             return f"# {path} (lines {start_line}-{end_line})\n\n{section}"
         except ValueError as e:
             log["error"] = str(e)
-            return f"Error: {e}"
+            raise ToolError(str(e))
 
 
 @mcp.tool(
@@ -556,14 +566,19 @@ async def search_docs(query: str, max_results: int = 5):
     Finds sections containing the query and returns previews with
     get_section() call hints so you can read the full content.
 
+    Multi-word queries use AND logic: all terms must appear in the
+    section (not necessarily on the same line).
+
     Args:
-        query: Exact string to search for (case-insensitive).
+        query: Search terms (case-insensitive). Multi-word queries
+               match sections containing ALL terms.
         max_results: Maximum sections to return (default: 5)
 
     Returns matching sections ranked by relevance with get_section() hints.
     """
     with _timed_tool("search_docs", query=query, max_results=max_results) as log:
-        pattern = re.compile(re.escape(query), re.IGNORECASE)
+        tokens = query.strip().split()
+        patterns = [re.compile(re.escape(t), re.IGNORECASE) for t in tokens]
         section_hits = []
 
         for rel_path in DOCS.keys():
@@ -580,8 +595,9 @@ async def search_docs(query: str, max_results: int = 5):
                 if header_match:
                     # Close previous section — check for hits
                     section_lines = lines[current_start:i]
-                    match_count = sum(1 for l in section_lines if pattern.search(l))
-                    if match_count and current_header != "(preamble)":
+                    section_text = "\n".join(section_lines)
+                    if all(p.search(section_text) for p in patterns) and current_header != "(preamble)":
+                        match_count = sum(sum(1 for p in patterns if p.search(l)) for l in section_lines)
                         section_hits.append({
                             "file": rel_path,
                             "header": current_header,
@@ -596,8 +612,9 @@ async def search_docs(query: str, max_results: int = 5):
 
             # Final section
             section_lines = lines[current_start:]
-            match_count = sum(1 for l in section_lines if pattern.search(l))
-            if match_count and current_header != "(preamble)":
+            section_text = "\n".join(section_lines)
+            if all(p.search(section_text) for p in patterns) and current_header != "(preamble)":
+                match_count = sum(sum(1 for p in patterns if p.search(l)) for l in section_lines)
                 section_hits.append({
                     "file": rel_path,
                     "header": current_header,
@@ -687,6 +704,10 @@ async def validate_function(fn_name: str) -> ValidationResult:
         if fn_name in PINE_V6_TOPLEVEL:
             return ValidationResult(valid=True, type="toplevel", function=fn_name)
 
+        # Check known invalid/renamed functions before generic fallback
+        if fn_name in KNOWN_REPLACEMENTS:
+            return ValidationResult(valid=False, type=None, function=fn_name, suggestion=KNOWN_REPLACEMENTS[fn_name])
+
         if "." in fn_name:
             ns = fn_name.rpartition(".")[0]
             if ns in PINE_V6_NAMESPACES:
@@ -738,6 +759,16 @@ async def resolve_topic(query: str) -> ResolveResult:
 
         log["matches_found"] = len(path_scores)
 
+        if not path_scores:
+            # Fallback: scan docs for an exact substring match before returning empty
+            fallback_pattern = re.compile(re.escape(query), re.IGNORECASE)
+            for rel_path in DOCS:
+                doc_lines = _get_doc_lines(rel_path)
+                if doc_lines and any(fallback_pattern.search(l) for l in doc_lines):
+                    path_scores[rel_path] = [query]
+                    break
+            log["fallback_used"] = bool(path_scores)
+
         if not path_scores:
             return ResolveResult(
                 matches=[],
@@ -930,6 +961,31 @@ async def metrics(request):
     return Response(generate_latest(METRICS_REGISTRY), media_type=CONTENT_TYPE_LATEST)
 
 
+class _AcceptNormalizer:
+    """Normalize Accept header on /mcp to prevent 406 Not Acceptable.
+
+    Workaround for modelcontextprotocol/python-sdk#2349 — the MCP SDK
+    requires both application/json AND text/event-stream in the Accept
+    header (even in SSE mode), but Anthropic's MCP proxy and other
+    clients send them separately per request type. Stamp the combined
+    value on /mcp only; SSE paths (/sse, /messages) pass through.
+    """
+    def __init__(self, app, mcp_path: str = "/mcp"):
+        self.app = app
+        self._mcp_path = mcp_path.rstrip("/")
+
+    async def __call__(self, scope, receive, send):
+        if scope.get("type") == "http" and scope.get("path", "").rstrip("/") == self._mcp_path:
+            headers = [
+                (b"accept", b"application/json, text/event-stream")
+                if name.lower() == b"accept"
+                else (name, value)
+                for name, value in scope.get("headers", [])
+            ]
+            scope = {**scope, "headers": headers}
+        await self.app(scope, receive, send)
+
+
 def main():
     """Entry point for the CLI."""
     import argparse
@@ -965,7 +1021,14 @@ def main():
             else:
                 await streamable_app(scope, receive, send)
 
-        config = uvicorn.Config(app, host=args.host, port=args.port, log_level="info")
+        config = uvicorn.Config(
+            _AcceptNormalizer(app),
+            host=args.host,
+            port=args.port,
+            log_level="info",
+            forwarded_allow_ips="*",
+            proxy_headers=True,
+        )
         server = uvicorn.Server(config)
         asyncio.run(server.serve())
     else: