@tyvm/knowhow 0.0.159 → 0.0.160

package/tests/clients/reasoningEffort.test.ts

@@ -1,140 +0,0 @@
-import { GenericAnthropicClient } from "../../src/clients/anthropic";
-import { GenericGeminiClient } from "../../src/clients/gemini";
-import { GenericOpenAiClient } from "../../src/clients/openai";
-import { AnthropicModels } from "../../src/clients/pricing";
-import {
-  REASONING_EFFORTS,
-  ReasoningEffort,
-  isReasoningEffort,
-} from "../../src/clients/types";
-import {
-  GoogleThinkingBudgetModels,
-  GoogleThinkingLevelModels,
-  Models,
-} from "../../src/types";
-import { GenericXAIClient } from "../../src/clients/xai";
-
-const options = (model: string, reasoning_effort: ReasoningEffort) => ({
-  model,
-  reasoning_effort,
-  messages: [],
-});
-
-describe("reasoning effort values", () => {
-  it("accepts the complete provider-neutral set and rejects other values", () => {
-    expect(REASONING_EFFORTS).toEqual([
-      "none",
-      "minimal",
-      "low",
-      "medium",
-      "high",
-      "xhigh",
-      "max",
-    ]);
-    REASONING_EFFORTS.forEach((effort) => expect(isReasoningEffort(effort)).toBe(true));
-    expect(isReasoningEffort("extra-high")).toBe(false);
-    expect(isReasoningEffort(undefined)).toBe(false);
-  });
-
-  it("maps provider-neutral values to OpenAI legacy and Responses spellings", () => {
-    const client = new GenericOpenAiClient("fake-key");
-    expect(client.resolveReasoningEffort(options("unknown", "none"))).toBe("low");
-    expect(client.resolveReasoningEffort(options("unknown", "minimal"))).toBe("low");
-    expect(client.resolveReasoningEffort(options("unknown", "xhigh"))).toBe("high");
-    expect(client.resolveReasoningEffort(options("unknown", "max"))).toBe("high");
-    expect(client.resolveReasoningEffortForModel(options("unknown", "max"))).toBe("xhigh");
-  });
-
-  it("explicitly disables reasoning in Responses API requests", async () => {
-    const client = new GenericOpenAiClient("fake-key");
-    const create = jest.fn().mockResolvedValue({
-      id: "resp_disabled",
-      status: "completed",
-      output: [{ type: "message", content: [{ type: "output_text", text: "summary" }] }],
-      usage: { input_tokens: 1, output_tokens: 1 },
-    });
-    (client.client as any).responses.create = create;
-
-    await client.createChatResponse({
-      model: Models.openai.GPT_56_Sol,
-      messages: [{ role: "user", content: "summarize" }],
-      reasoning_effort: "none",
-    });
-
-    expect(create).toHaveBeenCalledWith(
-      expect.objectContaining({ reasoning: { effort: "none" } })
-    );
-  });
-
-  it("diagnoses empty Responses output but not valid tool-only output", async () => {
-    const client = new GenericOpenAiClient("fake-key");
-    const warn = jest.spyOn(console, "warn").mockImplementation(() => undefined);
-    const create = jest.fn()
-      .mockResolvedValueOnce({
-        id: "resp_empty",
-        status: "incomplete",
-        incomplete_details: { reason: "max_output_tokens" },
-        output: [{ type: "reasoning", summary: [] }],
-        usage: { input_tokens: 2, output_tokens: 3 },
-      })
-      .mockResolvedValueOnce({
-        id: "resp_tool",
-        status: "completed",
-        output: [{
-          type: "function_call", call_id: "call_1", name: "readFile", arguments: "{}",
-        }],
-        usage: { input_tokens: 2, output_tokens: 3 },
-      });
-    (client.client as any).responses.create = create;
-    const request = {
-      model: Models.openai.GPT_56_Sol,
-      messages: [{ role: "user" as const, content: "continue" }],
-    };
-
-    await client.createChatResponse(request);
-    expect(warn).toHaveBeenCalledWith(
-      "[Responses API] Response contained no text or tool calls",
-      expect.objectContaining({ responseId: "resp_empty", status: "incomplete" })
-    );
-    warn.mockClear();
-    await client.createChatResponse(request);
-    expect(warn).not.toHaveBeenCalled();
-    warn.mockRestore();
-  });
-
-  it("maps all values to valid Gemini thinking levels and budgets", () => {
-    const client = new GenericGeminiClient("fake-key");
-    const levelModel = GoogleThinkingLevelModels.find(
-      (model) => model !== Models.google.Gemini_37_Flash
-    )!;
-    const budgetModel = GoogleThinkingBudgetModels[0];
-
-    expect(client.buildThinkingConfig(options(levelModel, "none"))).toEqual({ thinkingLevel: "minimal" });
-    expect(client.buildThinkingConfig(options(levelModel, "minimal"))).toEqual({ thinkingLevel: "minimal" });
-    expect(client.buildThinkingConfig(options(levelModel, "xhigh"))).toEqual({ thinkingLevel: "high" });
-    expect(client.buildThinkingConfig(options(levelModel, "max"))).toEqual({ thinkingLevel: "high" });
-    expect(client.buildThinkingConfig(options(Models.google.Gemini_37_Flash, "minimal"))).toEqual({ thinkingLevel: "low" });
-    expect(client.buildThinkingConfig(options(budgetModel, "none"))).toEqual({ thinkingBudget: 0 });
-    expect(client.buildThinkingConfig(options(budgetModel, "minimal"))).toEqual({ thinkingBudget: 512 });
-    expect(client.buildThinkingConfig(options(budgetModel, "medium"))).toEqual({ thinkingBudget: 8192 });
-    expect(client.buildThinkingConfig(options(budgetModel, "max"))).toEqual({ thinkingBudget: -1 });
-  });
-
-  it("maps provider-neutral values to Anthropic output effort", () => {
-    const client = new GenericAnthropicClient("fake-key");
-    expect(client.getEffortOutputConfig(AnthropicModels.Sonnet5, "none")).toBeUndefined();
-    expect(client.getEffortOutputConfig(AnthropicModels.Sonnet5, "minimal")).toEqual({ effort: "low" });
-    expect(client.getEffortOutputConfig(AnthropicModels.Sonnet5, "xhigh")).toEqual({ effort: "xhigh" });
-    expect(client.getEffortOutputConfig(AnthropicModels.Sonnet5, "max")).toEqual({ effort: "max" });
-    expect(client.getEffortOutputConfig("unsupported", "high")).toBeUndefined();
-  });
-
-  it("maps provider-neutral values to xAI legacy effort", () => {
-    const client = new GenericXAIClient("fake-key");
-    expect(client.resolveLegacyReasoningEffort("none")).toBeUndefined();
-    expect(client.resolveLegacyReasoningEffort("minimal")).toBe("low");
-    expect(client.resolveLegacyReasoningEffort("medium")).toBe("medium");
-    expect(client.resolveLegacyReasoningEffort("xhigh")).toBe("high");
-    expect(client.resolveLegacyReasoningEffort("max")).toBe("high");
-  });
-});

package/tests/compressor/bigstring.test.ts

@@ -1,450 +0,0 @@
-import { readFile } from "../../src/utils";
-import { TokenCompressor, KeyInfo } from "../../src/processors/TokenCompressor";
-import { services } from "../../src/services";
-import { Message } from "../../src/clients/types";
-
-describe("TokenCompressor - Large File Test", () => {
-  let tokenCompressor: TokenCompressor;
-  const bigstringPath = "tests/compressor/bigstring.txt";
-  const jsonPath = "tests/compressor/githubjson.txt";
-
-  beforeAll(() => {
-    const { Tools } = services();
-    tokenCompressor = new TokenCompressor(Tools);
-  });
-
-  afterEach(() => {
-    tokenCompressor.clearStorage();
-  });
-
-  test("should compress large file contents and allow retrieval via chunks", async () => {
-    // Load the large file
-    const fileBuffer = await readFile(bigstringPath);
-    const fileContents = fileBuffer.toString();
-
-    console.log(`Original file size: ${fileContents.length} characters`);
-    console.log(`Estimated tokens: ${Math.ceil(fileContents.length / 4)}`);
-
-    // Compress the content
-    const compressed = tokenCompressor.compressContent(
-      fileContents,
-      bigstringPath
-    );
-
-    console.log(`Compressed result: ${compressed}`);
-
-    // Verify that compression occurred
-    expect(compressed).toContain("[COMPRESSED_STRING");
-    expect(compressed).toContain("Key:");
-    expect(compressed).toContain("chunks]");
-    expect(compressed.length).toBeLessThan(fileContents.length);
-
-    // Extract the key from the compressed string
-    const keyMatch = compressed.match(/Key: (compressed_[a-z0-9_]+)/);
-    expect(keyMatch).not.toBeNull();
-    const firstKey = keyMatch![1];
-
-    // Retrieve the first chunk
-    const firstChunk = tokenCompressor.retrieveString(firstKey);
-    expect(firstChunk).toBeTruthy();
-    expect(firstChunk.length).toBeGreaterThan(0);
-
-    // Verify the first chunk contains the beginning of the original content
-    expect(fileContents.startsWith(firstChunk.split("[NEXT_CHUNK_KEY:")[0]));
-
-    // Follow the chain to retrieve all chunks
-    const currentChunk = firstChunk;
-    let reconstructed = "";
-    let chunkCount = 0;
-    const maxChunks = 100; // Safety limit
-
-    while (currentChunk && chunkCount < maxChunks) {
-      chunkCount++;
-
-      const nextKeyMatch = currentChunk.match(/\[NEXT_CHUNK_KEY: ([^\]]+)\]/);
-      if (nextKeyMatch) {
-        // Remove the NEXT_CHUNK_KEY marker and add content
-        const nextKey = nextKeyMatch[1];
-        const retrieved = await tokenCompressor.retrieveString(nextKey);
-        console.log(
-          `Retrieved chunk ${chunkCount} with key: ${nextKey}, length: ${retrieved.length}`
-        );
-      } else {
-        // Last chunk
-        reconstructed += currentChunk;
-        break;
-      }
-    }
-
-    console.log(`Retrieved ${chunkCount} chunks`);
-    console.log(`Reconstructed size: ${reconstructed.length} characters`);
-
-    // Verify the reconstructed content matches the original
-    // expect(reconstructed).toBe(fileContents);
-    expect(chunkCount).toBeGreaterThan(1); // Should have multiple chunks for a large file
-  });
-
-  test("should handle compression threshold correctly", async () => {
-    const fileBuffer = await readFile(bigstringPath);
-    const fileContents = fileBuffer.toString();
-
-    // Test that it compresses when above threshold
-    const estimatedTokens = Math.ceil(fileContents.length / 4);
-    expect(estimatedTokens).toBeGreaterThan(4000); // Default threshold
-
-    const compressed = tokenCompressor.compressContent(fileContents);
-
-    // Should be compressed
-    expect(compressed).toContain("[COMPRESSED_STRING");
-
-    // The compression should result in a much smaller representation
-    const compressionRatio = compressed.length / fileContents.length;
-    expect(compressionRatio).toBeLessThan(0.01); // Less than 1% of original size
-  });
-
-  test("should handle json compression", async () => {
-    const fileBuffer = await readFile(jsonPath);
-    const fileContents = fileBuffer.toString();
-
-    // Test that it compresses when above threshold
-    const estimatedTokens = Math.ceil(fileContents.length / 4);
-    expect(estimatedTokens).toBeGreaterThan(4000); // Default threshold
-
-    const compressed = tokenCompressor.compressContent(fileContents);
-
-    console.log({ compressedJson: compressed });
-
-    // Should return JSON object (not a string marker) for MCP format
-    const compressedObj = JSON.parse(compressed);
-    expect(compressedObj._mcp_format).toBe(true);
-    expect(compressedObj._data).toBeDefined();
-    expect(compressedObj._schema_key).toBeDefined();
-
-    // The compression should result in a smaller representation
-    const compressionRatio = compressed.length / fileContents.length;
-    expect(compressionRatio).toBeLessThan(0.5); // Less than 50% of original size
-  });
-
-  test("should detect json", async () => {
-    const fileBuffer = await readFile(jsonPath);
-    const fileContents = fileBuffer.toString();
-
-    const compressed = tokenCompressor.tryParseJson(fileContents);
-
-    expect(compressed).toBeTruthy();
-  });
-
-  test("should analyze githubjson.txt compression with key chain utilities", async () => {
-    const fileBuffer = await readFile(jsonPath);
-    const fileContents = fileBuffer.toString();
-
-    console.log(`\n=== GitHub JSON Compression Analysis ===`);
-    console.log(`Original file size: ${fileContents.length} characters`);
-    console.log(`Estimated tokens: ${Math.ceil(fileContents.length / 4)}`);
-
-    // Compress the content
-    const compressed = tokenCompressor.compressContent(fileContents);
-
-    console.log(
-      `\nCompressed representation length: ${compressed.length} characters`
-    );
-    console.log(`\nCompressed output:\n${compressed.substring(0, 500)}...\n`);
-
-    // Extract root keys from the compressed output
-    const rootKeys = tokenCompressor.extractKeys(compressed);
-    console.log(`\nRoot keys found: ${rootKeys.length}`);
-    console.log(`Keys: ${rootKeys.join(", ")}\n`);
-
-    expect(rootKeys.length).toBeGreaterThan(0);
-
-    // Analyze the key chain for each root key
-    for (const rootKey of rootKeys) {
-      console.log(`\n--- Analyzing key chain for: ${rootKey} ---`);
-      const keyChain = tokenCompressor.getKeyChain(rootKey);
-
-      console.log(`Total keys in chain: ${keyChain.length}`);
-
-      // Group keys by depth
-      const byDepth = keyChain.reduce((acc, info) => {
-        if (!acc[info.depth]) {
-          acc[info.depth] = [];
-        }
-        acc[info.depth].push(info);
-        return acc;
-      }, {} as Record<number, KeyInfo[]>);
-
-      console.log(`\nKeys by depth:`);
-      for (const [depth, infos] of Object.entries(byDepth)) {
-        console.log(`  Depth ${depth}: ${infos.length} keys`);
-        for (const info of infos) {
-          console.log(`    - ${info.key}:`);
-          console.log(
-            `        Size: ${info.size} chars (${info.tokens} tokens)`
-          );
-          console.log(`        Type: ${info.type}`);
-          console.log(`        Child keys: ${info.childKeys.length}`);
-          if (info.nextChunkKey) {
-            console.log(`        Next chunk: ${info.nextChunkKey}`);
-          }
-        }
-      }
-
-      // Verify chunk sizes are reasonable
-      for (const info of keyChain) {
-        const data = await tokenCompressor.retrieveString(info.key);
-        console.log({ data });
-
-        // Chunks should not be unnecessarily small (at least 25% of threshold unless it's the last chunk)
-        // Allow smaller chunks for next_chunk types at the end of chains
-        if (info.type === "child") {
-          expect(info.tokens).toBeGreaterThan(1000); // Child chunks should be substantial
-        }
-
-        // Chunks should not exceed the max tokens significantly
-        expect(info.tokens).toBeLessThan(tokenCompressor.maxTokens * 1.5);
-      }
-
-      // Calculate total size
-      const totalSize = keyChain.reduce((sum, info) => sum + info.size, 0);
-      const totalTokens = keyChain.reduce((sum, info) => sum + info.tokens, 0);
-      console.log(
-        `\nTotal stored size: ${totalSize} characters (${totalTokens} tokens)`
-      );
-      console.log(
-        `Storage efficiency: Stored ${keyChain.length} chunks for original content`
-      );
-    }
-
-    // Verify storage state
-    console.log(`\n=== Storage State ===`);
-    console.log(`Total keys in storage: ${tokenCompressor.getStorageSize()}`);
-    console.log(
-      `All storage keys: ${tokenCompressor.getStorageKeys().join(", ")}`
-    );
-
-    expect(tokenCompressor.getStorageSize()).toBeGreaterThan(0);
-
-    // Test extractKeys utility - with new schema format, compressed is a JSON object
-    // Parse it to find actual data keys
-    const parsedCompressed = tokenCompressor.tryParseJson(compressed);
-    expect(parsedCompressed).toBeTruthy();
-
-    // Look for keys in the parsed structure
-    let dataKeys: string[] = [];
-    if (parsedCompressed && typeof parsedCompressed === "object") {
-      // Extract all keys from the object recursively
-      const extractAllKeys = (obj: any): string[] => {
-        const keys: string[] = [];
-        const str = JSON.stringify(obj);
-        keys.push(...tokenCompressor.extractKeys(str));
-        return keys;
-      };
-      dataKeys = extractAllKeys(parsedCompressed);
-    }
-
-    console.log(`\nKeys found in compressed structure: ${dataKeys.length}`);
-    const storedContent =
-      dataKeys.length > 0 ? tokenCompressor.retrieveString(dataKeys[0]) : null;
-
-    const embeddedKeys = tokenCompressor.extractKeys(storedContent!);
-    console.log(
-      `\nKeys embedded in first stored chunk: ${embeddedKeys.length}`
-    );
-    if (embeddedKeys.length > 0) {
-      console.log(`Embedded keys: ${embeddedKeys.join(", ")}`);
-    }
-  });
-
-  test("should demonstrate improved compression with schema and low-signal detection", async () => {
-    const fileBuffer = await readFile(jsonPath);
-    const fileContents = fileBuffer.toString();
-
-    console.log(`\n=== Compression Performance Analysis ===`);
-    console.log(`Original file size: ${fileContents.length} characters`);
-    console.log(`Estimated tokens: ${Math.ceil(fileContents.length / 4)}`);
-
-    // Compress the content
-    const compressed = tokenCompressor.compressContent(fileContents);
-
-    console.log(`\nCompressed representation: ${compressed.length} characters`);
-    const compressionRatio = (
-      (1 - compressed.length / fileContents.length) *
-      100
-    ).toFixed(2);
-    console.log(`Compression ratio: ${compressionRatio}%`);
-
-    // Parse the compressed output
-    console.log(
-      `\nCompressed output preview:\n${compressed.substring(0, 800)}`
-    );
-
-    const compressedObj = JSON.parse(compressed);
-
-    console.log(`\n=== Schema Information ===`);
-    if (compressedObj._schema_key) {
-      console.log(`Schema key found: ${compressedObj._schema_key}`);
-      // The _schema_key already includes the full key with _schema suffix
-      const schema = tokenCompressor.retrieveString(compressedObj._schema_key);
-      if (schema) {
-        console.log(`Schema structure:`);
-        console.log(JSON.stringify(schema, null, 2).substring(0, 1000));
-      }
-    }
-
-    console.log(`\n=== Low-Signal Property Compression ===`);
-    // Check first item in the data array for compression metadata
-    if (compressedObj._data && Array.isArray(compressedObj._data)) {
-      const firstItem = compressedObj._data[0];
-      if (
-        typeof firstItem === "string" &&
-        firstItem.includes("COMPRESSED_JSON")
-      ) {
-        // Extract the key and retrieve the object
-        const keyMatch = firstItem.match(/Key: (compressed_[a-z0-9_]+)/);
-        if (keyMatch) {
-          const itemKey = keyMatch[1];
-          const itemData = tokenCompressor.retrieveString(itemKey);
-          if (itemData) {
-            const item = JSON.parse(itemData);
-            const innerItem = JSON.parse(item.text);
-
-            // Check for compressed properties in array items
-            if (Array.isArray(innerItem)) {
-              const sampleItem = innerItem[0];
-              if (sampleItem._compressed_properties_key) {
-                console.log(
-                  `Sample item has ${sampleItem._compressed_property_names.length} compressed properties`
-                );
-                console.log(
-                  `Compressed properties: ${sampleItem._compressed_property_names
-                    .slice(0, 10)
-                    .join(", ")}...`
-                );
-                console.log(`Info: ${sampleItem._compression_info}`);
-
-                // Retrieve the compressed properties
-                const compressedProps = tokenCompressor.getCompressedProperties(
-                  sampleItem._compressed_properties_key
-                );
-                if (compressedProps) {
-                  const propNames = Object.keys(
-                    compressedProps.compressed_properties
-                  );
-                  console.log(
-                    `\nRetrieved ${propNames.length} compressed properties successfully`
-                  );
-                  console.log(
-                    `Sample compressed property keys: ${propNames
-                      .slice(0, 5)
-                      .join(", ")}`
-                  );
-                }
-              }
-            }
-          }
-        }
-      }
-    }
-
-    console.log(`\n=== Storage Analysis ===`);
-    const totalKeys = tokenCompressor.getStorageSize();
-    console.log(`Total storage keys: ${totalKeys}`);
-
-    // Calculate total stored size
-    let totalStoredSize = 0;
-    for (const key of tokenCompressor.getStorageKeys()) {
-      const data = tokenCompressor.retrieveString(key);
-      if (data) {
-        totalStoredSize += data.length;
-      }
-    }
-
-    console.log(`Total stored data: ${totalStoredSize} characters`);
-    console.log(`Compressed representation: ${compressed.length} characters`);
-    console.log(
-      `Storage overhead ratio: ${(
-        (totalStoredSize / fileContents.length) *
-        100
-      ).toFixed(2)}%`
-    );
-    console.log(`Initial view compression: ${compressionRatio}%`);
-
-    // Verify we achieved good compression (showing first instance for usability)
-    expect(parseFloat(compressionRatio)).toBeGreaterThan(60); // At least 60% compression
-
-    // Verify schema is accessible
-    if (compressedObj._schema_key) {
-      const schema = tokenCompressor.retrieveString(compressedObj._schema_key);
-      expect(schema).toBeTruthy();
-    }
-
-    // Verify low-signal compression happened
-    expect(totalKeys).toBeGreaterThan(5); // Should have multiple compression keys
-
-    console.log(`\n✓ Compression improvements verified!`);
-  });
-
-  test("should not double-compress already compressed content", async () => {
-    // Load the githubjson.txt which contains a Message object
-    const fileBuffer = await readFile(jsonPath);
-    const fileContents = fileBuffer.toString();
-
-    console.log(`\n=== Testing Double Compression Prevention ===`);
-    console.log(`Original file size: ${fileContents.length} characters`);
-
-    // Parse the message from the file
-    const message = JSON.parse(fileContents) as Message;
-
-    // Create a copy of the message to compress
-    const messageToCompress: Message = JSON.parse(JSON.stringify(message));
-
-    // First compression
-    await tokenCompressor.compressMessage(messageToCompress);
-
-    const firstCompression = (messageToCompress.content as any[])[0].text;
-    console.log(`\nAfter first compression: ${firstCompression.length} characters`);
-    console.log(`First compression preview:\n${firstCompression.substring(0, 300)}...`);
-
-    // Parse the first compression result
-    const firstParsed = tokenCompressor.tryParseJson(firstCompression);
-    expect(firstParsed).toBeTruthy();
-    expect(firstParsed._schema_key).toBeDefined();
-    expect(firstParsed.data).toBeDefined();
-
-    // Second compression - this should NOT compress again
-    await tokenCompressor.compressMessage(messageToCompress);
-
-    const secondCompression = (messageToCompress.content as any[])[0].text;
-    console.log(`\nAfter second compression: ${secondCompression.length} characters`);
-    console.log(`Second compression preview:\n${secondCompression.substring(0, 300)}...`);
-
-    // Parse the second compression result
-    const secondParsed = tokenCompressor.tryParseJson(secondCompression);
-    expect(secondParsed).toBeTruthy();
-    
-    console.log(`\nSecond parsed structure keys:`, Object.keys(secondParsed));
-    console.log(`Full structure:`, JSON.stringify(secondParsed, null, 2).substring(0, 500));
-    
-    // Check if we get the over-compressed bug: only metadata, no data
-    const isOverCompressed = 
-      Object.keys(secondParsed).length === 3 &&
-      secondParsed._mcp_format === true &&
-      secondParsed._raw_structure !== undefined &&
-      secondParsed._schema_key !== undefined;
-    
-    console.log(`\nIs over-compressed: ${isOverCompressed}`);
-
-    // After the fix, second compression should be identical to first (no re-compression)
-    expect(secondCompression).toBe(firstCompression);
-    
-    // Both should have the same structure with data
-    expect(secondParsed).toEqual(firstParsed);
-    expect(secondParsed.data).toBeDefined();
-    expect(Array.isArray(secondParsed.data)).toBe(true);
-    
-    // Should NOT be over-compressed
-    expect(isOverCompressed).toBe(false);
-    
-    console.log(`\n✓ Double compression prevention verified!`);
-  });
-});