kcode-pi 0.1.30 → 0.1.34

package/extensions/kingdee-harness.ts

@@ -365,6 +365,16 @@ export default function (pi: ExtensionAPI) {
 			return { block: true, reason };
 		}
 
+		// Redirect read tool to kd_doc_read for office document formats
+		if (event.toolName === "read" && path) {
+			const ext = path.toLowerCase().split(".").pop();
+			if (ext === "pdf" || ext === "docx" || ext === "doc" || ext === "xlsx" || ext === "xls" || ext === "csv") {
+				const reason = `read 工具无法解析 .${ext} 二进制格式。下一步：改用 kd_doc_read 工具读取此文件，参数 path="${path}"。`;
+				if (ctx.hasUI) ctx.ui.notify(reason, "info");
+				return { block: true, reason };
+			}
+		}
+
 		const subagentRole = isSubagentChild() ? subagentRoleFromEnv() : undefined;
 		if (subagentRole) {
 			const run = readActiveRun(ctx.cwd);

package/extensions/kingdee-header.ts

@@ -4,6 +4,96 @@ import { readActiveRun } from "../src/harness/state.ts";
 import type { ActiveRun, GateResult } from "../src/harness/types.ts";
 import { formatProductProfile } from "../src/product/profile.ts";
 
+/** ANSI escape sequence pattern: CSI, OSC, APC. */
+const ANSI_RE = /\x1b\[[0-9;]*[A-Za-z]|\x1b\][^\x07]*\x07|\x1b_[^\x1b]*\x1b\\/g;
+
+/** Visible width of a string (strips ANSI codes; CJK/wide chars = 2 columns). */
+function visibleWidth(str: string): number {
+	let width = 0;
+	const clean = str.replace(ANSI_RE, "");
+	for (const ch of clean) {
+		const code = ch.codePointAt(0)!;
+		width +=
+			code >= 0x1100 &&
+			!(code >= 0x00a0 && code <= 0x00ff) &&
+			((code >= 0x1100 && code <= 0x115f) ||
+				(code >= 0x2329 && code <= 0x232a) ||
+				(code >= 0x2e80 && code <= 0x303e) ||
+				(code >= 0x3040 && code <= 0x3247) ||
+				(code >= 0x3250 && code <= 0x4dbf) ||
+				(code >= 0x4e00 && code <= 0xa4c6) ||
+				(code >= 0xa960 && code <= 0xa97c) ||
+				(code >= 0xac00 && code <= 0xd7a3) ||
+				(code >= 0xf900 && code <= 0xfaff) ||
+				(code >= 0xfe10 && code <= 0xfe19) ||
+				(code >= 0xfe30 && code <= 0xfe6b) ||
+				(code >= 0xff01 && code <= 0xff60) ||
+				(code >= 0xffe0 && code <= 0xffe6) ||
+				(code >= 0x1f300 && code <= 0x1f9ff) ||
+				(code >= 0x20000 && code <= 0x2fffd))
+				? 2
+				: 1;
+	}
+	return width;
+}
+
+/**
+ * If the line's visible width exceeds `maxWidth`, truncate visible characters
+ * and append `>` so the result fits. Preserves ANSI codes in the kept portion
+ * and appends SGR reset before the `>`. No padding — pi-tui only requires
+ * visibleWidth <= width.
+ */
+export function clipLine(text: string, maxWidth: number): string {
+	if (maxWidth <= 0) return "";
+	const vw = visibleWidth(text);
+	if (vw <= maxWidth) return text;
+
+	const targetW = maxWidth - 1; // reserve 1 col for ">"
+	let result = "";
+	let visibleSoFar = 0;
+	let i = 0;
+
+	while (i < text.length && visibleSoFar < targetW) {
+		// Preserve ANSI escape sequences
+		if (text[i] === "\x1b") {
+			const m = text.slice(i).match(/^\x1b(?:\[[0-9;]*[A-Za-z]|\][^\x07]*\x07|_[^\x1b]*\x1b\\)/);
+			if (m) {
+				result += m[0];
+				i += m[0].length;
+				continue;
+			}
+		}
+
+		const ch = text[i];
+		const code = ch.codePointAt(0)!;
+		const wide =
+			(code >= 0x1100 && code <= 0x115f) ||
+			(code >= 0x2329 && code <= 0x232a) ||
+			(code >= 0x2e80 && code <= 0x303e) ||
+			(code >= 0x3040 && code <= 0x3247) ||
+			(code >= 0x3250 && code <= 0x4dbf) ||
+			(code >= 0x4e00 && code <= 0xa4c6) ||
+			(code >= 0xa960 && code <= 0xa97c) ||
+			(code >= 0xac00 && code <= 0xd7a3) ||
+			(code >= 0xf900 && code <= 0xfaff) ||
+			(code >= 0xfe10 && code <= 0xfe19) ||
+			(code >= 0xfe30 && code <= 0xfe6b) ||
+			(code >= 0xff01 && code <= 0xff60) ||
+			(code >= 0xffe0 && code <= 0xffe6) ||
+			(code >= 0x1f300 && code <= 0x1f9ff) ||
+			(code >= 0x20000 && code <= 0x2fffd);
+		const charW = wide ? 2 : 1;
+
+		if (visibleSoFar + charW > targetW) break;
+
+		result += ch;
+		visibleSoFar += charW;
+		i++;
+	}
+
+	return result + "\x1b[0m>";
+}
+
 function formatProduct(run: ActiveRun | undefined): string {
 	if (!run) return "未选择";
 	if (run.profile?.product === "unknown") return "未确认";
@@ -43,12 +133,6 @@ function riskColor(risk: string): "error" | "warning" | "muted" | "success" {
 	return "success";
 }
 
-function padOrTrim(text: string, width: number): string {
-	if (width <= 0) return "";
-	if (text.length > width) return text.slice(0, Math.max(0, width - 1)) + ">";
-	return text + " ".repeat(width - text.length);
-}
-
 function logoLines(theme: Theme): string[] {
 	const accent = (text: string) => theme.fg("accent", text);
 	const muted = (text: string) => theme.fg("muted", text);
@@ -87,9 +171,9 @@ export default function (pi: ExtensionAPI) {
 
 					return [
 						"",
-						...logoLines(theme).map((line) => padOrTrim(line, width)),
-						padOrTrim(status, width),
-						padOrTrim(theme.fg("dim", `run：${runId}`), width),
+						...logoLines(theme).map((line) => clipLine(line, width)),
+						clipLine(status, width),
+						clipLine(theme.fg("dim", `run：${runId}`), width),
 						"",
 					];
 				},

package/extensions/kingdee-tools.ts

@@ -1,6 +1,6 @@
-import { dirname, join } from "node:path";
+import { dirname, join, extname } from "node:path";
 import { fileURLToPath } from "node:url";
-import { readFileSync } from "node:fs";
+import { readFileSync, readFile as fsReadFile } from "node:fs";
 import { Type } from "@earendil-works/pi-ai";
 import { defineTool, type ExtensionAPI } from "@earendil-works/pi-coding-agent";
 import { formatSearchResults, formatTableSchema } from "../src/knowledge/format.ts";
@@ -430,6 +430,133 @@ const kdDebugTool = defineTool({
 	},
 });
 
+const kdDocReadTool = defineTool({
+	name: "kd_doc_read",
+	label: "KD 文档读取",
+	description:
+		"读取 PDF、Word (.docx/.doc)、Excel (.xlsx/.xls) 或 CSV 文件并提取文本内容。对于 .pdf、.docx、.doc、.xlsx、.xls、.csv 文件，必须使用此工具而非 read 工具，因为 read 无法解析这些二进制格式。PDF 目前只支持可复制文本抽取；扫描型 PDF 需要另行提供图片或 OCR 结果。",
+	parameters: Type.Object({
+		path: Type.String({ description: "文档文件路径，支持 .pdf、.docx、.doc、.xlsx、.xls、.csv。" }),
+		sheet: Type.Optional(Type.String({ description: "Excel 工作表名或序号（从 1 开始）。默认读取第一个工作表。" })),
+		maxRows: Type.Optional(Type.Number({ description: "Excel/CSV 最大返回行数。默认 200。" })),
+		maxPages: Type.Optional(Type.Number({ description: "PDF 最大提取页数。默认 50。" })),
+	}),
+
+	async execute(_toolCallId, params, _signal, _onUpdate, ctx) {
+		const filePath = resolveWorkspacePath(ctx.cwd, params.path);
+		const ext = extname(filePath).toLowerCase();
+
+		try {
+			let text: string;
+
+			if (ext === ".pdf") {
+				text = await extractPdf(filePath, params.maxPages ?? 50);
+			} else if (ext === ".docx") {
+				text = await extractDocx(filePath);
+			} else if (ext === ".doc") {
+				text = await extractDoc(filePath);
+			} else if (ext === ".xlsx" || ext === ".xls") {
+				text = extractXlsx(filePath, params.sheet, params.maxRows ?? 200);
+			} else if (ext === ".csv") {
+				text = extractCsv(filePath, params.maxRows ?? 200);
+			} else {
+				return {
+					content: [{ type: "text", text: `不支持的文件格式：${ext}。支持 .pdf、.docx、.doc、.xlsx、.xls、.csv。` }],
+					details: { error: "unsupported-format", ext },
+				};
+			}
+
+			return {
+				content: [{ type: "text", text }],
+				details: { path: params.path, format: ext },
+			};
+		} catch (error) {
+			return {
+				content: [{ type: "text", text: `读取文档失败：${error instanceof Error ? error.message : String(error)}` }],
+				details: { error: "doc-read-failed", path: params.path },
+			};
+		}
+	},
+});
+
+async function extractPdf(filePath: string, maxPages: number): Promise<string> {
+	const { PDFParse } = await import("pdf-parse");
+	const buffer = readFileSync(filePath);
+	const parser = new PDFParse({ data: new Uint8Array(buffer) });
+	const result = await parser.getText({ last: maxPages });
+	const pages = result.total;
+	const text = result.text.trim();
+	return `[PDF] ${pages} 页\n\n${text}`;
+}
+
+async function extractDocx(filePath: string): Promise<string> {
+	const mammoth = await import("mammoth");
+	const buffer = readFileSync(filePath);
+	const result = await mammoth.extractRawText({ buffer });
+	return `[Word .docx]\n\n${result.value.trim()}`;
+}
+
+async function extractDoc(filePath: string): Promise<string> {
+	const word = await import("word");
+	const doc = await word.readFile(filePath);
+	const text = word.to_text(doc);
+	return `[Word .doc]\n\n${text.trim()}`;
+}
+
+function extractXlsx(filePath: string, sheetNameOrIndex?: string, maxRows?: number): string {
+	const XLSX = require("xlsx");
+	const workbook = XLSX.readFile(filePath);
+	const sheetNames = workbook.SheetNames;
+
+	let sheetName: string;
+	if (sheetNameOrIndex) {
+		const idx = Number(sheetNameOrIndex);
+		if (!isNaN(idx) && idx >= 1 && idx <= sheetNames.length) {
+			sheetName = sheetNames[idx - 1];
+		} else if (sheetNames.includes(sheetNameOrIndex)) {
+			sheetName = sheetNameOrIndex;
+		} else {
+			return `[Excel] 工作表 "${sheetNameOrIndex}" 不存在。可用：${sheetNames.join(", ")}`;
+		}
+	} else {
+		sheetName = sheetNames[0];
+	}
+
+	const sheet = workbook.Sheets[sheetName];
+	const rows: Record<string, unknown>[] = XLSX.utils.sheet_to_json(sheet, { defval: "" });
+	const limit = Math.min(rows.length, maxRows ?? 200);
+
+	if (rows.length === 0) return `[Excel] 工作表 "${sheetName}" 为空。`;
+
+	const headers = Object.keys(rows[0]);
+	const lines: string[] = [
+		`[Excel] 工作表：${sheetName}（${sheetNames.length} 个：${sheetNames.join(", ")}）`,
+		`行数：${rows.length}${rows.length > limit ? `（显示前 ${limit} 行）` : ""}`,
+		"",
+		"| " + headers.join(" | ") + " |",
+		"| " + headers.map(() => "---").join(" | ") + " |",
+	];
+
+	for (let i = 0; i < limit; i++) {
+		const row = rows[i];
+		lines.push("| " + headers.map((h) => String(row[h] ?? "")).join(" | ") + " |");
+	}
+
+	return lines.join("\n");
+}
+
+function extractCsv(filePath: string, maxRows: number): string {
+	const XLSX = require("xlsx");
+	const workbook = XLSX.readFile(filePath, { raw: true });
+	const sheetName = workbook.SheetNames[0];
+	const sheet = workbook.Sheets[sheetName];
+	const csv = XLSX.utils.sheet_to_csv(sheet);
+	const lines = csv.split("\n");
+	const limit = Math.min(lines.length, maxRows + 1); // +1 for header
+	const truncated = lines.length > limit;
+	return `[CSV] ${lines.length - 1} 行${truncated ? `（显示前 ${limit - 1} 行）` : ""}\n\n${lines.slice(0, limit).join("\n")}`;
+}
+
 export default function (pi: ExtensionAPI) {
 	pi.registerTool(kdSearchTool);
 	pi.registerTool(kdTableTool);
@@ -441,6 +568,7 @@ export default function (pi: ExtensionAPI) {
 	pi.registerTool(kdKsqlLintTool);
 	pi.registerTool(kdBuildTool);
 	pi.registerTool(kdDebugTool);
+	pi.registerTool(kdDocReadTool);
 }
 
 function writeSdkSignatureEvidence(cwd: string, content: string): string | undefined {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "kcode-pi",
-  "version": "0.1.30",
+  "version": "0.1.34",
   "description": "面向金蝶开发的 Pi Coding Agent 启动器、工具包和 Harness 工作流",
   "type": "module",
   "private": false,
@@ -47,7 +47,11 @@
   },
   "dependencies": {
     "@earendil-works/pi-ai": "^0.78.1",
-    "@earendil-works/pi-coding-agent": "^0.78.1"
+    "@earendil-works/pi-coding-agent": "^0.78.1",
+    "mammoth": "^1.12.0",
+    "pdf-parse": "^2.4.5",
+    "word": "^0.4.0",
+    "xlsx": "^0.18.5"
   },
   "devDependencies": {
     "@types/node": "^25.9.2",

package/src/harness/gates.ts

@@ -16,6 +16,7 @@ import {
 	missingForTargetReason,
 	missingMarkerReason,
 	openQuestionsReason,
+	repairBlockedReason,
 	unknownProductReason,
 	unknownRiskReason,
 } from "./messages.ts";
@@ -71,6 +72,7 @@ function collectGateProblems(cwd: string, run: ActiveRun, phase: KdPhase, mode:
 	const stepProblem = inspectStepState(cwd, run, phase);
 	const evidenceProblem = mode === "inspect" ? inspectEvidence(cwd, run, phase) : undefined;
 	const questionProblem = inspectOpenQuestions(run);
+	const repairProblem = inspectRepairState(run);
 
 	if (flagshipPathProblem) {
 		reasonParts.push(flagshipPathProblem);
@@ -87,6 +89,9 @@ function collectGateProblems(cwd: string, run: ActiveRun, phase: KdPhase, mode:
 	if (questionProblem) {
 		reasonParts.push(questionProblem);
 	}
+	if (repairProblem) {
+		reasonParts.push(repairProblem);
+	}
 
 	if (mode === "enter") {
 		missing.push(...missingEvidenceForPhase(cwd, run, phase));
@@ -148,6 +153,11 @@ function inspectOpenQuestions(run: ActiveRun): string | undefined {
 	return openQuestionsReason(open);
 }
 
+function inspectRepairState(run: ActiveRun): string | undefined {
+	if (run.repair?.status !== "blocked") return undefined;
+	return repairBlockedReason(run.repair.attempts, run.repair.maxAttempts, run.repair.lastFailureEvidence);
+}
+
 function inspectStepState(cwd: string, run: ActiveRun, phase: KdPhase): string | undefined {
 	if (phase === "execute") {
 		const plan = readArtifact(cwd, run, "plan") ?? "";

package/src/harness/messages.ts

@@ -19,6 +19,11 @@ export function openQuestionsReason(questions: Array<{ id: string; question: str
 	].join("。");
 }
 
+export function repairBlockedReason(attempts: number, maxAttempts: number, evidence?: string): string {
+	const evidenceText = evidence ? `，最近失败证据：${evidence}` : "";
+	return `自动修复已停止：验证失败 ${attempts}/${maxAttempts}${evidenceText}。下一步：回到 plan 调整范围，或重新创建/回答修复问题选择“继续修复”。`;
+}
+
 export function missingMarkerReason(phase: KdPhase, missing: string[]): string {
 	return `${PHASE_ARTIFACTS[phase]} 缺少必需章节：${missing.join(", ")}。下一步：更新 ${PHASE_ARTIFACTS[phase]}，补齐这些章节并写入真实内容。`;
 }

package/src/harness/state.ts

@@ -123,6 +123,7 @@ export function answerQuestion(cwd: string, run: ActiveRun, id: string, answer:
 	question.status = "answered";
 	question.answer = answer.trim();
 	question.answeredAt = new Date().toISOString();
+	applyRepairQuestionAnswer(cwd, run, question);
 	run.gate = inspectGate(cwd, run);
 	writeActiveRun(cwd, run);
 	return question;
@@ -216,6 +217,62 @@ function writeRunState(cwd: string, run: ActiveRun): void {
 	writeFileSync(runStatePath(cwd, run), `${JSON.stringify(run, null, 2)}\n`, "utf8");
 }
 
+function applyRepairQuestionAnswer(cwd: string, run: ActiveRun, question: KdQuestion): void {
+	if (!isRepairLimitQuestion(question)) return;
+	const answer = question.answer?.trim() ?? "";
+	const now = new Date().toISOString();
+	const attempts = run.repair?.attempts ?? 0;
+	const maxAttempts = run.repair?.maxAttempts ?? 3;
+	const extendedMaxAttempts = attempts + maxAttempts;
+
+	if (/^继续修复\b|^继续|continue/i.test(answer)) {
+		run.phase = "execute";
+		ensureArtifact(cwd, run, "execute", defaultArtifactContent("execute", run.goal, run.profile));
+		run.repair = {
+			attempts,
+			maxAttempts: extendedMaxAttempts,
+			lastFailureEvidence: run.repair?.lastFailureEvidence,
+			lastFailureSignature: run.repair?.lastFailureSignature,
+			status: "repairing",
+			updatedAt: now,
+		};
+		return;
+	}
+
+	if (/^回到\s*plan\b|^回到计划|^返回\s*plan\b|^重新计划/i.test(answer)) {
+		run.phase = "plan";
+		ensureArtifact(cwd, run, "plan", defaultArtifactContent("plan", run.goal, run.profile));
+		run.repair = {
+			attempts,
+			maxAttempts: extendedMaxAttempts,
+			lastFailureEvidence: run.repair?.lastFailureEvidence,
+			lastFailureSignature: run.repair?.lastFailureSignature,
+			status: "idle",
+			updatedAt: now,
+		};
+		return;
+	}
+
+	if (/^停止|^stop\b/i.test(answer)) {
+		run.repair = {
+			attempts: run.repair?.attempts ?? maxAttempts,
+			maxAttempts,
+			lastFailureEvidence: run.repair?.lastFailureEvidence,
+			lastFailureSignature: run.repair?.lastFailureSignature,
+			status: "blocked",
+			updatedAt: now,
+		};
+	}
+}
+
+function isRepairLimitQuestion(question: KdQuestion): boolean {
+	return (
+		question.phase === "verify" &&
+		/验证失败已达到/.test(question.question) &&
+		(Array.isArray(question.choices) ? question.choices.includes("继续修复") : true)
+	);
+}
+
 function hydrateRun(parsed: ActiveRun): ActiveRun | undefined {
 	if (!parsed || typeof parsed !== "object") return undefined;
 	if (typeof parsed.id !== "string" || !parsed.id.trim()) return undefined;