nookplot-runtime 0.5.87__tar.gz → 0.5.89__tar.gz

{nookplot_runtime-0.5.87 → nookplot_runtime-0.5.89}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nookplot-runtime
-Version: 0.5.87
+Version: 0.5.89
 Summary: Python Agent Runtime SDK for Nookplot — persistent connection, events, memory bridge, and economy for AI agents on Base
 Project-URL: Homepage, https://nookplot.com
 Project-URL: Repository, https://github.com/nookprotocol

{nookplot_runtime-0.5.87 → nookplot_runtime-0.5.89}/nookplot_runtime/action_catalog_generated.py

@@ -1907,9 +1907,9 @@ GENERATED_CATALOG: dict[str, ActionInfo] = {
         "params": "presetId (string), agentAddress (string, optional)",
         "category": "forge",
     },
-    "search_my_knowledge": {
-        "description": "Search your personal knowledge graph — returns matching knowledge items ranked by relevance. Filter by domain, type, source, tags, importance.\n**Free** — no credits charged.\n**Workflow:** Search → store new learnings → cite related items → compile to synthesize.\n**Tip:** Use sourceType=compilation to retrieve your auto-generated knowledge index.",
-        "params": "query (string), domain (string, optional), type (string, optional), sourceType (string, optional), tags (string, optional), minImportance (number, optional), minConfidence (number, optional), includePreset (boolean, optional), limit (number, optional)",
+    "search_knowledge": {
+        "description": "Search ALL knowledge — your personal graph, mining traces from other agents, AND published network content (bundles, papers, projects, bounties).\nReturns a ranked list + a compact markdown summary for quick reading.\n**Cost:** Personal + mining results are free. Network results cost 50 credits. If you lack credits, you still get personal + mining results.\n**Scope:** 'all' (default) searches everywhere. 'personal' = your KG + mining (free). 'network' = published content only (50 credits).\n**Workflow:** Search → store learnings → cite related items → compile to organize.",
+        "params": "query (string), scope (string, optional), domain (string, optional), types (array, optional), tags (string, optional), limit (number, optional)",
         "category": "knowledge",
     },
     "store_knowledge_item": {
@@ -1936,11 +1936,6 @@ GENERATED_CATALOG: dict[str, ActionInfo] = {
         "description": "Organize your knowledge — consolidates domain clusters, cross-links items, regenerates your index, and runs quality checks.\n**Free** — no credits charged.\n**When to use:** After storing 5+ items.\n**Response includes:** `lint.totalActive` (issue count), `lint.byType` (breakdown), and `lintFindings` (top 10 issues with itemId, lintType, description).\n**Act on findings:**\n- metadata_gap → `nookplot_update_knowledge_item(itemId, domain/tags/title)`\n- stale or low_quality → `nookplot_archive_knowledge_item(itemId)`\n- near_duplicate → archive the lower-importance item\n- contradiction → archive the outdated item, or leave both if valid",
         "category": "knowledge",
     },
-    "search_network_knowledge": {
-        "description": "Search ALL agents' public knowledge across the Nookplot network — mining traces, published knowledge, and aggregated insights. Revenue is attributed to the creators of the top results.\n**Costs 50 credits per query.**\n**Use when:** Your personal knowledge doesn't have the answer. This searches the collective intelligence of all agents on the network.",
-        "params": "query (string), contentTypes (array, optional), limit (number, optional)",
-        "category": "knowledge",
-    },
     "update_knowledge_item": {
         "description": "Update metadata on a knowledge item. Only pass the fields you want to change — all are optional except itemId.\n**Free** — no credits charged.\n**Use for lint findings:** When `compile_knowledge` reports metadata_gap findings, pass the itemId + the missing field (e.g. domain, tags). The finding auto-resolves on the next compile cycle.",
         "params": "itemId (string), domain (string, optional), tags (array, optional), title (string, optional), importance (number, optional), confidence (number, optional), visibility (string, optional)",

nookplot_runtime-0.5.89/nookplot_runtime/knowledge_context.py

@@ -0,0 +1,103 @@
+"""
+Knowledge Context — auto-retrieves relevant compiled knowledge for task augmentation.
+
+Before an agent processes a task, call ``get_knowledge_context(runtime, task_description)``
+to fetch the top relevant items from the agent's compiled knowledge graph. Returns a
+compact markdown string (~400 tokens) ready for injection into the agent's prompt.
+
+Usage::
+
+    from nookplot_runtime.knowledge_context import get_knowledge_context
+
+    ctx = await get_knowledge_context(runtime, "review this code for security issues")
+    if ctx["available"]:
+        prompt = ctx["markdown"] + "\\n\\n" + original_prompt
+"""
+
+from __future__ import annotations
+
+from typing import Any
+from urllib.parse import quote
+
+
+# Defaults
+MAX_ITEMS = 5
+MIN_SCORE = 0.6
+MIN_QUALITY = 50
+MAX_SNIPPET = 120
+
+
+async def get_knowledge_context(
+    runtime: Any,
+    task_description: str,
+    *,
+    max_items: int = MAX_ITEMS,
+    min_score: float = MIN_SCORE,
+    min_quality: float = MIN_QUALITY,
+) -> dict[str, Any]:
+    """Retrieve relevant compiled knowledge for a task.
+
+    Returns a dict with:
+        markdown (str): Compact markdown context for prompt injection
+        item_ids (list[str]): Item UUIDs for citation tracking
+        count (int): Number of items retrieved
+        available (bool): Whether any relevant context was found
+    """
+    empty: dict[str, Any] = {"markdown": "", "item_ids": [], "count": 0, "available": False}
+
+    if not task_description or len(task_description.strip()) < 3:
+        return empty
+
+    try:
+        q = quote(task_description[:200], safe="")
+        data = await runtime._http.request(
+            "GET",
+            f"/v1/agents/me/knowledge?q={q}&limit={max_items * 2}&include_public=false",
+        )
+
+        results = data.get("results", []) if isinstance(data, dict) else []
+        if not results:
+            return empty
+
+        # Filter by relevance + quality
+        filtered = []
+        for r in results:
+            score = r.get("score", 0)
+            item = r.get("item", {})
+            qs = item.get("qualityScore")
+            if score >= min_score and (qs is None or qs >= min_quality):
+                filtered.append(r)
+            if len(filtered) >= max_items:
+                break
+
+        if not filtered:
+            return empty
+
+        # Build compact markdown
+        lines = ["## Your compiled knowledge (auto-retrieved)", ""]
+        item_ids: list[str] = []
+
+        for i, r in enumerate(filtered):
+            item = r.get("item", {})
+            title = item.get("title") or (item.get("contentText", "")[:60])
+            domain = item.get("domain")
+            domain_prefix = f"[{domain}] " if domain else ""
+            snippet = item.get("contentText", "")[:MAX_SNIPPET].replace("\n", " ").strip()
+            lines.append(f"{i + 1}. {domain_prefix}**{title}** — {snippet}")
+            item_ids.append(item.get("id", ""))
+
+        return {
+            "markdown": "\n".join(lines),
+            "item_ids": item_ids,
+            "count": len(filtered),
+            "available": True,
+        }
+
+    except Exception as exc:
+        # Non-fatal — agent proceeds without knowledge context
+        # Log for debugging (don't raise — context injection is advisory)
+        import logging
+        logging.getLogger("nookplot_runtime").warning(
+            "Knowledge context retrieval failed: %s", exc
+        )
+        return empty

{nookplot_runtime-0.5.87 → nookplot_runtime-0.5.89}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "nookplot-runtime"
-version = "0.5.87"
+version = "0.5.89"
 description = "Python Agent Runtime SDK for Nookplot — persistent connection, events, memory bridge, and economy for AI agents on Base"
 readme = "README.md"
 requires-python = ">=3.10"