chainlesschain 0.45.12 → 0.45.19

package/src/lib/prompt-compressor.js

@@ -1,481 +1,10 @@
-/**
- * CLI Prompt Compressor — 5 strategies for context window management.
- *
- * Strategies:
- *   1. deduplication  — Remove duplicate/similar messages (Jaccard similarity)
- *   2. truncation     — Keep most recent N messages
- *   3. summarization  — LLM-generated summary of old history
- *   4. snipCompact    — Remove stale tool results and processed markers
- *   5. contextCollapse — Fold consecutive same-type messages into summaries
- *
- * Feature-flag gated: PROMPT_COMPRESSOR, CONTEXT_SNIP, CONTEXT_COLLAPSE
- */
-
-import { createHash } from "node:crypto";
-import { feature } from "./feature-flags.js";
-
-// ── Token estimation ────────────────────────────────────────────────────
-
-/**
- * Estimate token count for a string.
- * Chinese: ~1.5 chars/token, English: ~4 chars/token.
- */
-export function estimateTokens(text) {
-  if (!text) return 0;
-  const chineseChars = (text.match(/[\u4e00-\u9fa5]/g) || []).length;
-  const otherChars = text.length - chineseChars;
-  return Math.ceil(chineseChars / 1.5 + otherChars / 4);
-}
-
-/**
- * Estimate total tokens for a messages array.
- */
-export function estimateMessagesTokens(messages) {
-  return messages.reduce((sum, msg) => {
-    const content =
-      typeof msg.content === "string"
-        ? msg.content
-        : JSON.stringify(msg.content || "");
-    return sum + estimateTokens(content);
-  }, 0);
-}
-
-// ── Similarity ──────────────────────────────────────────────────────────
-
-function jaccardSimilarity(str1, str2) {
-  if (!str1 || !str2) return 0;
-  if (str1 === str2) return 1;
-  const tokens1 = new Set(str1.split(""));
-  const tokens2 = new Set(str2.split(""));
-  let intersection = 0;
-  for (const t of tokens1) {
-    if (tokens2.has(t)) intersection++;
-  }
-  return intersection / (tokens1.size + tokens2.size - intersection);
-}
-
-function getContent(msg) {
-  return typeof msg.content === "string"
-    ? msg.content
-    : JSON.stringify(msg.content || "");
-}
-
-// ── Provider context window registry ───────────────────────────────────
-
-/**
- * Known context window sizes (in tokens) per model/provider.
- * Used by adaptive compression to auto-tune maxTokens.
- */
-export const CONTEXT_WINDOWS = {
-  // Ollama local models
-  "qwen2.5:7b": 32768,
-  "qwen2.5:14b": 32768,
-  "qwen2.5-coder:14b": 32768,
-  "qwen2:7b": 32768,
-  "llama3:8b": 8192,
-  "mistral:7b": 32768,
-  "codellama:7b": 16384,
-  // OpenAI
-  "gpt-4o": 128000,
-  "gpt-4o-mini": 128000,
-  "gpt-4-turbo": 128000,
-  "gpt-3.5-turbo": 16385,
-  o1: 200000,
-  // Anthropic
-  "claude-opus-4-6": 200000,
-  "claude-sonnet-4-6": 200000,
-  "claude-haiku-4-5-20251001": 200000,
-  // DeepSeek
-  "deepseek-chat": 64000,
-  "deepseek-coder": 64000,
-  "deepseek-reasoner": 64000,
-  // DashScope
-  "qwen-turbo": 131072,
-  "qwen-plus": 131072,
-  "qwen-max": 32768,
-  // Gemini
-  "gemini-2.0-flash": 1048576,
-  "gemini-2.0-pro": 1048576,
-  "gemini-1.5-flash": 1048576,
-  // Kimi
-  "moonshot-v1-auto": 131072,
-  "moonshot-v1-8k": 8192,
-  "moonshot-v1-32k": 32768,
-  "moonshot-v1-128k": 131072,
-  // Volcengine
-  "doubao-seed-1-6-251015": 32768,
-  // Provider-level defaults (fallback when model not listed)
-  _provider_defaults: {
-    ollama: 32768,
-    openai: 128000,
-    anthropic: 200000,
-    deepseek: 64000,
-    dashscope: 131072,
-    gemini: 1048576,
-    kimi: 131072,
-    volcengine: 32768,
-    minimax: 32768,
-    mistral: 32768,
-  },
-};
-
-/**
- * Get context window size for a model/provider combination.
- * @param {string} [model] — Model name
- * @param {string} [provider] — Provider name
- * @returns {number} Context window in tokens
- */
-export function getContextWindow(model, provider) {
-  if (model && CONTEXT_WINDOWS[model]) {
-    return CONTEXT_WINDOWS[model];
-  }
-  if (provider && CONTEXT_WINDOWS._provider_defaults[provider]) {
-    return CONTEXT_WINDOWS._provider_defaults[provider];
-  }
-  return 32768; // Conservative default
-}
-
-/**
- * Calculate adaptive compression thresholds based on context window.
- *
- * Strategy:
- * - maxTokens = 60% of context window (reserve 40% for system prompt + response)
- * - maxMessages scales with context: small ctx → 15, large ctx → 50
- * - For very large contexts (>128k), enable less aggressive compression
- *
- * @param {number} contextWindow — Context window in tokens
- * @returns {{ maxMessages: number, maxTokens: number, aggressive: boolean }}
- */
-export function adaptiveThresholds(contextWindow) {
-  const maxTokens = Math.floor(contextWindow * 0.6);
-  // Scale messages: 15 for 8k, 20 for 32k, 30 for 128k, 50 for 200k+
-  const maxMessages = Math.min(
-    50,
-    Math.max(15, Math.floor(10 + Math.log2(contextWindow / 1024) * 5)),
-  );
-  // Aggressive compression only for small context windows
-  const aggressive = contextWindow < 32768;
-
-  return { maxMessages, maxTokens, aggressive };
-}
-
-// ── PromptCompressor class ──────────────────────────────────────────────
-
-export class PromptCompressor {
-  /**
-   * @param {object} options
-   * @param {number} [options.maxMessages=20] — Max messages before truncation
-   * @param {number} [options.maxTokens=8000] — Token threshold for auto-compact
-   * @param {number} [options.similarityThreshold=0.9] — Dedup similarity threshold
-   * @param {Function} [options.llmQuery] — async (prompt) => string, for summarization
-   * @param {string} [options.model] — Model name (for adaptive thresholds)
-   * @param {string} [options.provider] — Provider name (for adaptive thresholds)
-   */
-  constructor(options = {}) {
-    // If model/provider supplied and no explicit maxMessages/maxTokens, auto-adapt
-    if (
-      (options.model || options.provider) &&
-      !options.maxMessages &&
-      !options.maxTokens
-    ) {
-      const ctxWindow = getContextWindow(options.model, options.provider);
-      const adaptive = adaptiveThresholds(ctxWindow);
-      this.maxMessages = adaptive.maxMessages;
-      this.maxTokens = adaptive.maxTokens;
-      this._adaptive = true;
-      this._contextWindow = ctxWindow;
-    } else {
-      this.maxMessages = options.maxMessages || 20;
-      this.maxTokens = options.maxTokens || 8000;
-      this._adaptive = false;
-      this._contextWindow = null;
-    }
-    this.similarityThreshold = options.similarityThreshold || 0.9;
-    this.llmQuery = options.llmQuery || null;
-  }
-
-  /**
-   * Reconfigure thresholds for a new model/provider (e.g. after model switch).
-   * Only updates if no explicit overrides were set in constructor.
-   */
-  adaptToModel(model, provider) {
-    const ctxWindow = getContextWindow(model, provider);
-    const adaptive = adaptiveThresholds(ctxWindow);
-    this.maxMessages = adaptive.maxMessages;
-    this.maxTokens = adaptive.maxTokens;
-    this._adaptive = true;
-    this._contextWindow = ctxWindow;
-  }
-
-  /**
-   * Run all enabled compression strategies on messages.
-   * Returns { messages, stats }.
-   */
-  async compress(messages, options = {}) {
-    if (!Array.isArray(messages) || messages.length <= 2) {
-      return {
-        messages: Array.isArray(messages) ? [...messages] : [],
-        stats: { strategy: "none", saved: 0 },
-      };
-    }
-
-    const originalTokens = estimateMessagesTokens(messages);
-    let result = [...messages];
-    const applied = [];
-
-    // Strategy 4: snipCompact (remove stale tool results)
-    if (feature("CONTEXT_SNIP")) {
-      const before = result.length;
-      result = this._snipCompact(result);
-      if (result.length < before) applied.push("snip");
-    }
-
-    // Strategy 1: deduplication
-    if (result.length > 3) {
-      const before = result.length;
-      result = this._deduplicate(result);
-      if (result.length < before) applied.push("dedup");
-    }
-
-    // Strategy 5: contextCollapse (fold consecutive tool results)
-    if (feature("CONTEXT_COLLAPSE") && result.length > 6) {
-      const before = result.length;
-      result = this._contextCollapse(result);
-      if (result.length < before) applied.push("collapse");
-    }
-
-    // Strategy 2: truncation
-    if (result.length > this.maxMessages) {
-      result = this._truncate(result);
-      applied.push("truncate");
-    }
-
-    // Strategy 3: summarization (only if still over token limit)
-    const currentTokens = estimateMessagesTokens(result);
-    if (this.llmQuery && currentTokens > this.maxTokens && result.length > 4) {
-      try {
-        result = await this._summarize(result);
-        applied.push("summarize");
-      } catch (_err) {
-        // Summarization failed — continue with what we have
-      }
-    }
-
-    const compressedTokens = estimateMessagesTokens(result);
-    return {
-      messages: result,
-      stats: {
-        strategy: applied.join("+") || "none",
-        originalMessages: messages.length,
-        compressedMessages: result.length,
-        originalTokens,
-        compressedTokens,
-        saved: originalTokens - compressedTokens,
-        ratio: originalTokens > 0 ? compressedTokens / originalTokens : 1,
-      },
-    };
-  }
-
-  /**
-   * Check if auto-compact should trigger.
-   */
-  shouldAutoCompact(messages) {
-    return (
-      messages.length > this.maxMessages ||
-      estimateMessagesTokens(messages) > this.maxTokens
-    );
-  }
-
-  // ── Strategy 1: Deduplication ───────────────────────────────────────
-
-  _deduplicate(messages) {
-    const system = messages.filter((m) => m.role === "system");
-    const last = [...messages].reverse().find((m) => m.role === "user");
-    const rest = messages.filter((m) => m.role !== "system" && m !== last);
-
-    const seen = new Map();
-    const deduped = [];
-
-    for (const msg of rest) {
-      const content = getContent(msg);
-      const hash = createHash("md5").update(content).digest("hex");
-
-      if (seen.has(hash)) continue;
-
-      let isDup = false;
-      for (const [, existing] of seen) {
-        if (
-          jaccardSimilarity(content, getContent(existing)) >=
-          this.similarityThreshold
-        ) {
-          isDup = true;
-          break;
-        }
-      }
-
-      if (!isDup) {
-        seen.set(hash, msg);
-        deduped.push(msg);
-      }
-    }
-
-    const result = [...system, ...deduped];
-    if (last && !result.includes(last)) result.push(last);
-    return result;
-  }
-
-  // ── Strategy 2: Truncation ──────────────────────────────────────────
-
-  _truncate(messages) {
-    const system = messages.filter((m) => m.role === "system");
-    const last = [...messages].reverse().find((m) => m.role === "user");
-    const rest = messages.filter((m) => m.role !== "system" && m !== last);
-
-    let slots = this.maxMessages - system.length;
-    if (last) slots -= 1;
-
-    const recent = rest.slice(-Math.max(slots, 1));
-    const result = [...system, ...recent];
-    if (last && !result.includes(last)) result.push(last);
-    return result;
-  }
-
-  // ── Strategy 3: Summarization ───────────────────────────────────────
-
-  async _summarize(messages) {
-    const system = messages.filter((m) => m.role === "system");
-    const last = [...messages].reverse().find((m) => m.role === "user");
-    const toSummarize = messages.filter(
-      (m) => m.role !== "system" && m !== last,
-    );
-
-    if (toSummarize.length < 3) return messages;
-
-    const historyText = toSummarize
-      .map((m) => `${m.role}: ${getContent(m).slice(0, 500)}`)
-      .join("\n");
-
-    const summary = await this.llmQuery(
-      `Summarize this conversation history concisely, preserving key facts and decisions:\n\n${historyText}\n\nSummary:`,
-    );
-
-    if (!summary) return messages;
-
-    const result = [
-      ...system,
-      { role: "system", content: `[Conversation Summary]\n${summary}` },
-    ];
-    if (last) result.push(last);
-    return result;
-  }
-
-  // ── Strategy 4: Snip Compact ────────────────────────────────────────
-  // Removes stale markers: processed tool results older than recent window,
-  // empty assistant messages, and system messages with [PROCESSED] tags.
-
-  _snipCompact(messages) {
-    if (messages.length <= 4) return messages;
-
-    // Keep system[0] + last 4 messages untouched
-    const head = messages.slice(0, 1);
-    const middle = messages.slice(1, -4);
-    const tail = messages.slice(-4);
-
-    const snipped = middle.filter((msg) => {
-      const content = getContent(msg);
-
-      // Remove empty messages
-      if (!content || content.trim() === "") return false;
-
-      // Remove processed markers
-      if (content.includes("[PROCESSED]") || content.includes("[STALE]"))
-        return false;
-
-      // Remove tool_result messages that are just "ok" or empty JSON
-      if (msg.role === "tool") {
-        if (
-          content === "ok" ||
-          content === "{}" ||
-          content === "null" ||
-          content.length < 3
-        )
-          return false;
-      }
-
-      // Remove very short assistant acknowledgments in middle
-      if (msg.role === "assistant" && content.length < 10) return false;
-
-      return true;
-    });
-
-    return [...head, ...snipped, ...tail];
-  }
-
-  // ── Strategy 5: Context Collapse ────────────────────────────────────
-  // Folds consecutive tool_call + tool_result pairs into a single summary.
-
-  _contextCollapse(messages) {
-    if (messages.length <= 6) return messages;
-
-    const result = [];
-    let i = 0;
-
-    while (i < messages.length) {
-      const msg = messages[i];
-
-      // Detect consecutive tool sequences in the middle (not last 3)
-      if (
-        i > 0 &&
-        i < messages.length - 3 &&
-        msg.role === "assistant" &&
-        msg.tool_calls &&
-        msg.tool_calls.length > 0
-      ) {
-        // Collect this tool call + all following tool results
-        const toolGroup = [msg];
-        let j = i + 1;
-        while (j < messages.length - 3 && messages[j].role === "tool") {
-          toolGroup.push(messages[j]);
-          j++;
-        }
-
-        // Also collect next assistant with tool_calls (chained calls)
-        while (
-          j < messages.length - 3 &&
-          messages[j].role === "assistant" &&
-          messages[j].tool_calls
-        ) {
-          toolGroup.push(messages[j]);
-          j++;
-          while (j < messages.length - 3 && messages[j].role === "tool") {
-            toolGroup.push(messages[j]);
-            j++;
-          }
-        }
-
-        // Only collapse if we collected 3+ messages
-        if (toolGroup.length >= 3) {
-          const toolNames = toolGroup
-            .filter((m) => m.tool_calls)
-            .flatMap((m) =>
-              m.tool_calls.map((tc) => tc.function?.name || "tool"),
-            )
-            .filter(Boolean);
-          const uniqueTools = [...new Set(toolNames)];
-
-          result.push({
-            role: "system",
-            content: `[Collapsed ${toolGroup.length} tool messages: ${uniqueTools.join(", ")}]`,
-          });
-          i = j;
-          continue;
-        }
-      }
-
-      result.push(msg);
-      i++;
-    }
-
-    return result;
-  }
-}
+export {
+  estimateTokens,
+  estimateMessagesTokens,
+  CONTEXT_WINDOWS,
+  getContextWindow,
+  COMPRESSION_VARIANTS,
+  getCompressionVariant,
+  adaptiveThresholds,
+  PromptCompressor,
+} from "../harness/prompt-compressor.js";

package/src/lib/sub-agent-context.js

@@ -16,6 +16,9 @@ import {
   createWorktree,
   removeWorktree,
   isolateTask,
+  diffWorktree,
+  mergeWorktree,
+  worktreeLog,
 } from "./worktree-isolator.js";
 import { isGitRepo } from "./git-integration.js";
 
@@ -150,12 +153,29 @@ export class SubAgentContext {
         },
       );
 
-      // Annotate result with worktree info
+      // Annotate result with worktree info + diff preview
       if (result) {
+        let diffInfo = null;
+        let commits = [];
+        if (
+          hasChanges ||
+          worktreeLog(this._repoDir, `agent/${taskId}`).length > 0
+        ) {
+          try {
+            diffInfo = diffWorktree(this._repoDir, `agent/${taskId}`);
+            commits = worktreeLog(this._repoDir, `agent/${taskId}`);
+          } catch (_e) {
+            // Non-critical — diff preview is optional
+          }
+        }
         result.worktree = {
           branch,
           path: worktreePath,
           hasChanges,
+          diff: diffInfo,
+          commits,
+          merge: (options = {}) =>
+            mergeWorktree(this._repoDir, branch, options),
         };
       }
       return result;