open-local-audit 0.67.0 → 0.69.0

package/dist/start.js

@@ -0,0 +1,234 @@
+import { lstat } from "node:fs/promises";
+import { join, resolve } from "node:path";
+import { createInterface } from "node:readline/promises";
+const categories = [
+    { query: "dental", profile: "dental" },
+    { query: "restaurant", profile: "restaurant" },
+    { query: "beauty", profile: "beauty" },
+    { query: "hotel", profile: "hotel" },
+    { query: "gym", profile: "gym" }
+];
+const overtureCategoryPattern = /^[a-z][a-z0-9_]{0,80}$/i;
+function defaultPromptIo() {
+    const isTTY = Boolean(process.stdin.isTTY && process.stdout.isTTY);
+    const reader = isTTY ? createInterface({ input: process.stdin, output: process.stdout, terminal: true }) : undefined;
+    let closed = false;
+    const close = () => {
+        if (!closed) {
+            closed = true;
+            reader?.close();
+        }
+    };
+    reader?.on("SIGINT", close);
+    return {
+        isTTY,
+        question: async (prompt) => {
+            if (!reader || closed)
+                return null;
+            return new Promise((answer) => {
+                const finish = (value) => {
+                    reader.off("close", onClose);
+                    answer(value);
+                };
+                const onClose = () => finish(null);
+                reader.once("close", onClose);
+                reader.question(prompt).then((value) => finish(value), () => finish(null));
+            });
+        },
+        write: (message) => process.stdout.write(message),
+        close,
+        cwd: () => process.cwd(),
+        exists: async (path) => {
+            try {
+                await lstat(path);
+                return true;
+            }
+            catch (error) {
+                if (error && typeof error === "object" && "code" in error && error.code === "ENOENT")
+                    return false;
+                throw error;
+            }
+        },
+        now: () => new Date()
+    };
+}
+async function answer(io, prompt) {
+    try {
+        return await io.question(prompt);
+    }
+    catch {
+        return null;
+    }
+}
+function writeLine(io, message) {
+    io.write(`${message}\n`);
+}
+function slug(value) {
+    const normalized = value
+        .normalize("NFKD")
+        .replace(/\p{M}/gu, "")
+        .toLowerCase()
+        .replace(/[^a-z0-9]+/g, "-")
+        .replace(/^-+|-+$/g, "");
+    return normalized || "search";
+}
+function timestamp(value) {
+    const pad = (part) => String(part).padStart(2, "0");
+    return `${value.getUTCFullYear()}${pad(value.getUTCMonth() + 1)}${pad(value.getUTCDate())}-${pad(value.getUTCHours())}${pad(value.getUTCMinutes())}${pad(value.getUTCSeconds())}`;
+}
+async function promptCountry(io) {
+    for (;;) {
+        const value = await answer(io, "Country ISO code (two letters): ");
+        if (value === null)
+            return null;
+        const country = value.trim().toUpperCase();
+        if (/^[A-Z]{2}$/.test(country))
+            return country;
+        writeLine(io, "Country must use a two-letter ISO code, for example TR.");
+    }
+}
+async function promptCity(io) {
+    for (;;) {
+        const value = await answer(io, "City: ");
+        if (value === null)
+            return null;
+        const city = value.trim();
+        if (city)
+            return city;
+        writeLine(io, "City is required.");
+    }
+}
+async function promptCategory(io) {
+    writeLine(io, "Category: 1) dental  2) restaurant  3) beauty  4) hotel  5) gym  6) custom Overture category");
+    for (;;) {
+        const selected = await answer(io, "Choose a category [1-6]: ");
+        if (selected === null)
+            return null;
+        if (!/^[1-6]$/.test(selected.trim())) {
+            writeLine(io, "Choose a category number from 1 to 6.");
+            continue;
+        }
+        const index = Number(selected.trim()) - 1;
+        if (index < categories.length)
+            return categories[index];
+        for (;;) {
+            const custom = await answer(io, "Custom Overture category: ");
+            if (custom === null)
+                return null;
+            const query = custom.trim();
+            if (overtureCategoryPattern.test(query))
+                return { query, profile: "generic" };
+            writeLine(io, "Use an Overture category identifier beginning with a letter and containing only letters, digits, or underscores.");
+        }
+    }
+}
+async function promptWholeNumber(io, prompt, defaultValue, minimum, maximum, invalidMessage) {
+    for (;;) {
+        const input = await answer(io, `${prompt} [${defaultValue}]: `);
+        if (input === null)
+            return null;
+        const value = input.trim();
+        if (!value)
+            return defaultValue;
+        if (/^(?:0|[1-9]\d*)$/.test(value)) {
+            const parsed = Number(value);
+            if (Number.isSafeInteger(parsed) && parsed >= minimum && parsed <= maximum)
+                return parsed;
+        }
+        writeLine(io, invalidMessage);
+    }
+}
+async function uniqueDefaultDirectory(io, country, city, category) {
+    const base = resolve(io.cwd(), "reports", `${country.toLowerCase()}-${slug(city)}-${slug(category)}-${timestamp(io.now())}`);
+    let candidate = base;
+    let suffix = 2;
+    while (await io.exists(candidate)) {
+        candidate = `${base}-${suffix}`;
+        suffix += 1;
+    }
+    return candidate;
+}
+async function promptOutputDirectory(io, defaultDirectory) {
+    for (;;) {
+        const input = await answer(io, `Output directory [${defaultDirectory}]: `);
+        if (input === null)
+            return null;
+        const directory = input.trim() ? resolve(io.cwd(), input.trim()) : defaultDirectory;
+        if (!await io.exists(directory))
+            return directory;
+        writeLine(io, "Output path already exists. Choose a new directory so existing results are not overwritten.");
+    }
+}
+function showSummary(io, options) {
+    writeLine(io, "");
+    writeLine(io, "Discovery summary");
+    writeLine(io, `  Location: ${options.city}, ${options.country}`);
+    writeLine(io, `  Category: ${options.query}`);
+    writeLine(io, `  Candidate limit: ${options.limit}`);
+    writeLine(io, `  Website audit cap: ${options.maxAudits}`);
+    writeLine(io, `  Output: ${options.outDir}`);
+    writeLine(io, "  Free API-keyless Overture discovery; results are stored in local files.");
+}
+export async function collectStartOptions(io = defaultPromptIo()) {
+    try {
+        if (!io.isTTY) {
+            throw new Error("Interactive start requires a terminal. Use discover dental --city Istanbul --country TR --dry-run --export-csv leads.csv instead.");
+        }
+        const country = await promptCountry(io);
+        if (country === null)
+            return null;
+        const city = await promptCity(io);
+        if (city === null)
+            return null;
+        const category = await promptCategory(io);
+        if (category === null)
+            return null;
+        const limit = await promptWholeNumber(io, "Candidate count", 10, 1, 100, "Candidate count must be a whole number from 1 to 100.");
+        if (limit === null)
+            return null;
+        const maxAudits = await promptWholeNumber(io, "Website audit cap", Math.min(3, limit), 0, limit, `Website audit cap must be a whole number from 0 to ${limit}.`);
+        if (maxAudits === null)
+            return null;
+        const defaultDirectory = await uniqueDefaultDirectory(io, country, city, category.query);
+        let outDir = await promptOutputDirectory(io, defaultDirectory);
+        if (outDir === null)
+            return null;
+        for (;;) {
+            showSummary(io, { city, country, query: category.query, limit, maxAudits, outDir });
+            const confirmation = await answer(io, "Start? [Y/n]: ");
+            if (confirmation === null)
+                return null;
+            const normalized = confirmation.trim().toLowerCase();
+            if (!normalized || normalized === "y" || normalized === "yes") {
+                if (!await io.exists(outDir)) {
+                    return {
+                        provider: "overture",
+                        query: category.query,
+                        city,
+                        country,
+                        profile: category.profile,
+                        limit,
+                        maxAudits,
+                        concurrency: 3,
+                        dryRun: maxAudits === 0,
+                        outDir,
+                        exportCsv: join(outDir, "leads.csv"),
+                        summaryJson: join(outDir, "discovery-summary.json")
+                    };
+                }
+                writeLine(io, "Output path already exists. Choose a new directory so existing results are not overwritten.");
+                outDir = await promptOutputDirectory(io, defaultDirectory);
+                if (outDir === null)
+                    return null;
+                continue;
+            }
+            if (normalized === "n" || normalized === "no")
+                return null;
+            writeLine(io, "Please answer Y or n.");
+        }
+    }
+    finally {
+        io.close();
+    }
+}
+//# sourceMappingURL=start.js.map

package/dist/start.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"start.js","sourceRoot":"","sources":["../src/start.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,EAAE,MAAM,kBAAkB,CAAC;AACzC,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAC1C,OAAO,EAAE,eAAe,EAAE,MAAM,wBAAwB,CAAC;AAmBzD,MAAM,UAAU,GAAqB;IACnC,EAAE,KAAK,EAAE,QAAQ,EAAE,OAAO,EAAE,QAAQ,EAAE;IACtC,EAAE,KAAK,EAAE,YAAY,EAAE,OAAO,EAAE,YAAY,EAAE;IAC9C,EAAE,KAAK,EAAE,QAAQ,EAAE,OAAO,EAAE,QAAQ,EAAE;IACtC,EAAE,KAAK,EAAE,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE;IACpC,EAAE,KAAK,EAAE,KAAK,EAAE,OAAO,EAAE,KAAK,EAAE;CACjC,CAAC;AAEF,MAAM,uBAAuB,GAAG,yBAAyB,CAAC;AAE1D,SAAS,eAAe;IACtB,MAAM,KAAK,GAAG,OAAO,CAAC,OAAO,CAAC,KAAK,CAAC,KAAK,IAAI,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;IACnE,MAAM,MAAM,GAAG,KAAK,CAAC,CAAC,CAAC,eAAe,CAAC,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,OAAO,CAAC,MAAM,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;IACrH,IAAI,MAAM,GAAG,KAAK,CAAC;IAEnB,MAAM,KAAK,GAAG,GAAG,EAAE;QACjB,IAAI,CAAC,MAAM,EAAE,CAAC;YACZ,MAAM,GAAG,IAAI,CAAC;YACd,MAAM,EAAE,KAAK,EAAE,CAAC;QAClB,CAAC;IACH,CAAC,CAAC;IAEF,MAAM,EAAE,EAAE,CAAC,QAAQ,EAAE,KAAK,CAAC,CAAC;IAE5B,OAAO;QACL,KAAK;QACL,QAAQ,EAAE,KAAK,EAAE,MAAM,EAAE,EAAE;YACzB,IAAI,CAAC,MAAM,IAAI,MAAM;gBAAE,OAAO,IAAI,CAAC;YAEnC,OAAO,IAAI,OAAO,CAAgB,CAAC,MAAM,EAAE,EAAE;gBAC3C,MAAM,MAAM,GAAG,CAAC,KAAoB,EAAE,EAAE;oBACtC,MAAM,CAAC,GAAG,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;oBAC7B,MAAM,CAAC,KAAK,CAAC,CAAC;gBAChB,CAAC,CAAC;gBACF,MAAM,OAAO,GAAG,GAAG,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;gBACnC,MAAM,CAAC,IAAI,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC;gBAC9B,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,GAAG,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC;YAC7E,CAAC,CAAC,CAAC;QACL,CAAC;QACD,KAAK,EAAE,CAAC,OAAO,EAAE,EAAE,CAAC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,OAAO,CAAC;QACjD,KAAK;QACL,GAAG,EAAE,GAAG,EAAE,CAAC,OAAO,CAAC,GAAG,EAAE;QACxB,MAAM,EAAE,KAAK,EAAE,IAAI,EAAE,EAAE;YACrB,IAAI,CAAC;gBACH,MAAM,KAAK,CAAC,IAAI,CAAC,CAAC;gBAClB,OAAO,IAAI,CAAC;YACd,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,IAAI,KAAK,IAAI,OAAO,KAAK,KAAK,QAAQ,IAAI,MAAM,IAAI,KAAK,IAAI,KAAK,CAAC,IAAI,KAAK,QAAQ;oBAAE,OAAO,KAAK,CAAC;gBACnG,MAAM,KAAK,CAAC;YACd,CAAC;QACH,CAAC;QACD,GAAG,EAAE,GAAG,EAAE,CAAC,IAAI,IAAI,EAAE;KACtB,CAAC;AACJ,CAAC;AAED,KAAK,UAAU,MAAM,CAAC,EAAiB,EAAE,MAAc;IACrD,IAAI,CAAC;QACH,OAAO,MAAM,EAAE,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;IACnC,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAC;IACd,CAAC;AACH,CAAC;AAED,SAAS,SAAS,CAAC,EAAiB,EAAE,OAAe;IACnD,EAAE,CAAC,KAAK,CAAC,GAAG,OAAO,IAAI,CAAC,CAAC;AAC3B,CAAC;AAED,SAAS,IAAI,CAAC,KAAa;IACzB,MAAM,UAAU,GAAG,KAAK;SACrB,SAAS,CAAC,MAAM,CAAC;SACjB,OAAO,CAAC,SAAS,EAAE,EAAE,CAAC;SACtB,WAAW,EAAE;SACb,OAAO,CAAC,aAAa,EAAE,GAAG,CAAC;SAC3B,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC,CAAC;IAC3B,OAAO,UAAU,IAAI,QAAQ,CAAC;AAChC,CAAC;AAED,SAAS,SAAS,CAAC,KAAW;IAC5B,MAAM,GAAG,GAAG,CAAC,IAAY,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,QAAQ,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;IAC5D,OAAO,GAAG,KAAK,CAAC,cAAc,EAAE,GAAG,GAAG,CAAC,KAAK,CAAC,WAAW,EAAE,GAAG,CAAC,CAAC,GAAG,GAAG,CAAC,KAAK,CAAC,UAAU,EAAE,CAAC,IAAI,GAAG,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,GAAG,GAAG,CAAC,KAAK,CAAC,aAAa,EAAE,CAAC,GAAG,GAAG,CAAC,KAAK,CAAC,aAAa,EAAE,CAAC,EAAE,CAAC;AACpL,CAAC;AAED,KAAK,UAAU,aAAa,CAAC,EAAiB;IAC5C,SAAS,CAAC;QACR,MAAM,KAAK,GAAG,MAAM,MAAM,CAAC,EAAE,EAAE,kCAAkC,CAAC,CAAC;QACnE,IAAI,KAAK,KAAK,IAAI;YAAE,OAAO,IAAI,CAAC;QAChC,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;QAC3C,IAAI,YAAY,CAAC,IAAI,CAAC,OAAO,CAAC;YAAE,OAAO,OAAO,CAAC;QAC/C,SAAS,CAAC,EAAE,EAAE,yDAAyD,CAAC,CAAC;IAC3E,CAAC;AACH,CAAC;AAED,KAAK,UAAU,UAAU,CAAC,EAAiB;IACzC,SAAS,CAAC;QACR,MAAM,KAAK,GAAG,MAAM,MAAM,CAAC,EAAE,EAAE,QAAQ,CAAC,CAAC;QACzC,IAAI,KAAK,KAAK,IAAI;YAAE,OAAO,IAAI,CAAC;QAChC,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC;QAC1B,IAAI,IAAI;YAAE,OAAO,IAAI,CAAC;QACtB,SAAS,CAAC,EAAE,EAAE,mBAAmB,CAAC,CAAC;IACrC,CAAC;AACH,CAAC;AAED,KAAK,UAAU,cAAc,CAAC,EAAiB;IAC7C,SAAS,CAAC,EAAE,EAAE,8FAA8F,CAAC,CAAC;IAC9G,SAAS,CAAC;QACR,MAAM,QAAQ,GAAG,MAAM,MAAM,CAAC,EAAE,EAAE,2BAA2B,CAAC,CAAC;QAC/D,IAAI,QAAQ,KAAK,IAAI;YAAE,OAAO,IAAI,CAAC;QACnC,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC;YACrC,SAAS,CAAC,EAAE,EAAE,uCAAuC,CAAC,CAAC;YACvD,SAAS;QACX,CAAC;QAED,MAAM,KAAK,GAAG,MAAM,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC,GAAG,CAAC,CAAC;QAC1C,IAAI,KAAK,GAAG,UAAU,CAAC,MAAM;YAAE,OAAO,UAAU,CAAC,KAAK,CAAC,CAAC;QAExD,SAAS,CAAC;YACR,MAAM,MAAM,GAAG,MAAM,MAAM,CAAC,EAAE,EAAE,4BAA4B,CAAC,CAAC;YAC9D,IAAI,MAAM,KAAK,IAAI;gBAAE,OAAO,IAAI,CAAC;YACjC,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC;YAC5B,IAAI,uBAAuB,CAAC,IAAI,CAAC,KAAK,CAAC;gBAAE,OAAO,EAAE,KAAK,EAAE,OAAO,EAAE,SAAS,EAAE,CAAC;YAC9E,SAAS,CAAC,EAAE,EAAE,kHAAkH,CAAC,CAAC;QACpI,CAAC;IACH,CAAC;AACH,CAAC;AAED,KAAK,UAAU,iBAAiB,CAC9B,EAAiB,EACjB,MAAc,EACd,YAAoB,EACpB,OAAe,EACf,OAAe,EACf,cAAsB;IAEtB,SAAS,CAAC;QACR,MAAM,KAAK,GAAG,MAAM,MAAM,CAAC,EAAE,EAAE,GAAG,MAAM,KAAK,YAAY,KAAK,CAAC,CAAC;QAChE,IAAI,KAAK,KAAK,IAAI;YAAE,OAAO,IAAI,CAAC;QAChC,MAAM,KAAK,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC;QAC3B,IAAI,CAAC,KAAK;YAAE,OAAO,YAAY,CAAC;QAChC,IAAI,kBAAkB,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;YACnC,MAAM,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC;YAC7B,IAAI,MAAM,CAAC,aAAa,CAAC,MAAM,CAAC,IAAI,MAAM,IAAI,OAAO,IAAI,MAAM,IAAI,OAAO;gBAAE,OAAO,MAAM,CAAC;QAC5F,CAAC;QACD,SAAS,CAAC,EAAE,EAAE,cAAc,CAAC,CAAC;IAChC,CAAC;AACH,CAAC;AAED,KAAK,UAAU,sBAAsB,CAAC,EAAiB,EAAE,OAAe,EAAE,IAAY,EAAE,QAAgB;IACtG,MAAM,IAAI,GAAG,OAAO,CAAC,EAAE,CAAC,GAAG,EAAE,EAAE,SAAS,EAAE,GAAG,OAAO,CAAC,WAAW,EAAE,IAAI,IAAI,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,QAAQ,CAAC,IAAI,SAAS,CAAC,EAAE,CAAC,GAAG,EAAE,CAAC,EAAE,CAAC,CAAC;IAC7H,IAAI,SAAS,GAAG,IAAI,CAAC;IACrB,IAAI,MAAM,GAAG,CAAC,CAAC;IACf,OAAO,MAAM,EAAE,CAAC,MAAM,CAAC,SAAS,CAAC,EAAE,CAAC;QAClC,SAAS,GAAG,GAAG,IAAI,IAAI,MAAM,EAAE,CAAC;QAChC,MAAM,IAAI,CAAC,CAAC;IACd,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC;AAED,KAAK,UAAU,qBAAqB,CAAC,EAAiB,EAAE,gBAAwB;IAC9E,SAAS,CAAC;QACR,MAAM,KAAK,GAAG,MAAM,MAAM,CAAC,EAAE,EAAE,qBAAqB,gBAAgB,KAAK,CAAC,CAAC;QAC3E,IAAI,KAAK,KAAK,IAAI;YAAE,OAAO,IAAI,CAAC;QAChC,MAAM,SAAS,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,CAAC,GAAG,EAAE,EAAE,KAAK,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,gBAAgB,CAAC;QACpF,IAAI,CAAC,MAAM,EAAE,CAAC,MAAM,CAAC,SAAS,CAAC;YAAE,OAAO,SAAS,CAAC;QAClD,SAAS,CAAC,EAAE,EAAE,6FAA6F,CAAC,CAAC;IAC/G,CAAC;AACH,CAAC;AAED,SAAS,WAAW,CAAC,EAAiB,EAAE,OAAmG;IACzI,SAAS,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC;IAClB,SAAS,CAAC,EAAE,EAAE,mBAAmB,CAAC,CAAC;IACnC,SAAS,CAAC,EAAE,EAAE,eAAe,OAAO,CAAC,IAAI,KAAK,OAAO,CAAC,OAAO,EAAE,CAAC,CAAC;IACjE,SAAS,CAAC,EAAE,EAAE,eAAe,OAAO,CAAC,KAAK,EAAE,CAAC,CAAC;IAC9C,SAAS,CAAC,EAAE,EAAE,sBAAsB,OAAO,CAAC,KAAK,EAAE,CAAC,CAAC;IACrD,SAAS,CAAC,EAAE,EAAE,wBAAwB,OAAO,CAAC,SAAS,EAAE,CAAC,CAAC;IAC3D,SAAS,CAAC,EAAE,EAAE,aAAa,OAAO,CAAC,MAAM,EAAE,CAAC,CAAC;IAC7C,SAAS,CAAC,EAAE,EAAE,2EAA2E,CAAC,CAAC;AAC7F,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,mBAAmB,CAAC,KAAoB,eAAe,EAAE;IAC7E,IAAI,CAAC;QACH,IAAI,CAAC,EAAE,CAAC,KAAK,EAAE,CAAC;YACd,MAAM,IAAI,KAAK,CAAC,mIAAmI,CAAC,CAAC;QACvJ,CAAC;QAED,MAAM,OAAO,GAAG,MAAM,aAAa,CAAC,EAAE,CAAC,CAAC;QACxC,IAAI,OAAO,KAAK,IAAI;YAAE,OAAO,IAAI,CAAC;QAClC,MAAM,IAAI,GAAG,MAAM,UAAU,CAAC,EAAE,CAAC,CAAC;QAClC,IAAI,IAAI,KAAK,IAAI;YAAE,OAAO,IAAI,CAAC;QAC/B,MAAM,QAAQ,GAAG,MAAM,cAAc,CAAC,EAAE,CAAC,CAAC;QAC1C,IAAI,QAAQ,KAAK,IAAI;YAAE,OAAO,IAAI,CAAC;QACnC,MAAM,KAAK,GAAG,MAAM,iBAAiB,CAAC,EAAE,EAAE,iBAAiB,EAAE,EAAE,EAAE,CAAC,EAAE,GAAG,EAAE,uDAAuD,CAAC,CAAC;QAClI,IAAI,KAAK,KAAK,IAAI;YAAE,OAAO,IAAI,CAAC;QAChC,MAAM,SAAS,GAAG,MAAM,iBAAiB,CAAC,EAAE,EAAE,mBAAmB,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,KAAK,CAAC,EAAE,CAAC,EAAE,KAAK,EAAE,sDAAsD,KAAK,GAAG,CAAC,CAAC;QACjK,IAAI,SAAS,KAAK,IAAI;YAAE,OAAO,IAAI,CAAC;QAEpC,MAAM,gBAAgB,GAAG,MAAM,sBAAsB,CAAC,EAAE,EAAE,OAAO,EAAE,IAAI,EAAE,QAAQ,CAAC,KAAK,CAAC,CAAC;QACzF,IAAI,MAAM,GAAG,MAAM,qBAAqB,CAAC,EAAE,EAAE,gBAAgB,CAAC,CAAC;QAC/D,IAAI,MAAM,KAAK,IAAI;YAAE,OAAO,IAAI,CAAC;QAEjC,SAAS,CAAC;YACR,WAAW,CAAC,EAAE,EAAE,EAAE,IAAI,EAAE,OAAO,EAAE,KAAK,EAAE,QAAQ,CAAC,KAAK,EAAE,KAAK,EAAE,SAAS,EAAE,MAAM,EAAE,CAAC,CAAC;YACpF,MAAM,YAAY,GAAG,MAAM,MAAM,CAAC,EAAE,EAAE,gBAAgB,CAAC,CAAC;YACxD,IAAI,YAAY,KAAK,IAAI;gBAAE,OAAO,IAAI,CAAC;YACvC,MAAM,UAAU,GAAG,YAAY,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC;YACrD,IAAI,CAAC,UAAU,IAAI,UAAU,KAAK,GAAG,IAAI,UAAU,KAAK,KAAK,EAAE,CAAC;gBAC9D,IAAI,CAAC,MAAM,EAAE,CAAC,MAAM,CAAC,MAAM,CAAC,EAAE,CAAC;oBAC7B,OAAO;wBACL,QAAQ,EAAE,UAAU;wBACpB,KAAK,EAAE,QAAQ,CAAC,KAAK;wBACrB,IAAI;wBACJ,OAAO;wBACP,OAAO,EAAE,QAAQ,CAAC,OAAO;wBACzB,KAAK;wBACL,SAAS;wBACT,WAAW,EAAE,CAAC;wBACd,MAAM,EAAE,SAAS,KAAK,CAAC;wBACvB,MAAM;wBACN,SAAS,EAAE,IAAI,CAAC,MAAM,EAAE,WAAW,CAAC;wBACpC,WAAW,EAAE,IAAI,CAAC,MAAM,EAAE,wBAAwB,CAAC;qBACpD,CAAC;gBACJ,CAAC;gBACD,SAAS,CAAC,EAAE,EAAE,6FAA6F,CAAC,CAAC;gBAC7G,MAAM,GAAG,MAAM,qBAAqB,CAAC,EAAE,EAAE,gBAAgB,CAAC,CAAC;gBAC3D,IAAI,MAAM,KAAK,IAAI;oBAAE,OAAO,IAAI,CAAC;gBACjC,SAAS;YACX,CAAC;YACD,IAAI,UAAU,KAAK,GAAG,IAAI,UAAU,KAAK,IAAI;gBAAE,OAAO,IAAI,CAAC;YAC3D,SAAS,CAAC,EAAE,EAAE,uBAAuB,CAAC,CAAC;QACzC,CAAC;IACH,CAAC;YAAS,CAAC;QACT,EAAE,CAAC,KAAK,EAAE,CAAC;IACb,CAAC;AACH,CAAC"}

package/dist/website-enrichment.d.ts

@@ -0,0 +1,21 @@
+import type { PageSnapshot, PublicContact } from "./types.js";
+type Resolver = (hostname: string) => Promise<string[]>;
+export interface WebsiteEnrichmentResult {
+    status: "success" | "blocked" | "failed";
+    snapshot?: PageSnapshot;
+    contact?: PublicContact;
+    sourceUrls: string[];
+    pagesFetched: number;
+    durationMs: number;
+    error?: string;
+    warnings?: string[];
+}
+export interface WebsiteEnrichmentOptions {
+    timeoutMs?: number;
+    maxPages?: number;
+    fetch?: typeof fetch;
+    /** Test seam for deterministic DNS checks. Production uses Node's resolver. */
+    resolve?: Resolver;
+}
+export declare function enrichWebsite(url: string, options?: WebsiteEnrichmentOptions): Promise<WebsiteEnrichmentResult>;
+export {};

package/dist/website-enrichment.js

@@ -0,0 +1,400 @@
+import { lookup as dnsLookup } from "node:dns/promises";
+import http from "node:http";
+import https from "node:https";
+import { createRequire } from "node:module";
+import { isIP } from "node:net";
+import { load } from "cheerio";
+import { extractPublicContact } from "./contact.js";
+const USER_AGENT = "open-local-audit/0.1 (+https://github.com/Esquetta/open-local-audit)";
+const DEFAULT_TIMEOUT_MS = 10_000;
+const MAX_HTML_BYTES = 1_048_576;
+const MAX_REDIRECTS = 3;
+const require = createRequire(import.meta.url);
+const robotsParser = require("robots-parser");
+class BlockedError extends Error {
+}
+class TimeoutError extends Error {
+}
+function normalizeHeaders(headers) {
+    const output = {};
+    headers.forEach((value, key) => {
+        output[key] = value;
+    });
+    return output;
+}
+function sameBusinessHost(first, second) {
+    return first.replace(/^www\./i, "").toLowerCase() === second.replace(/^www\./i, "").toLowerCase();
+}
+function isPublicAddress(address) {
+    const normalized = address.toLowerCase().split("%")[0] ?? "";
+    if (isIP(normalized) === 4) {
+        const octets = normalized.split(".").map(Number);
+        const [first, second] = octets;
+        return !(first === 0 ||
+            first === 10 ||
+            first === 127 ||
+            first >= 224 ||
+            (first === 192 && second === 0) ||
+            (first === 198 && (second === 18 || second === 19)) ||
+            (first === 100 && second >= 64 && second <= 127) ||
+            (first === 169 && second === 254) ||
+            (first === 172 && second >= 16 && second <= 31) ||
+            (first === 192 && second === 168));
+    }
+    if (isIP(normalized) !== 6) {
+        return false;
+    }
+    // Only native global unicast; exclude mapped/translation/tunnel and documentation ranges.
+    return /^[23]/.test(normalized) && !/^2001:(?:0:|db8:)/.test(normalized) && !normalized.startsWith("2002:");
+}
+async function resolvePublicAddresses(hostname, resolve) {
+    hostname = hostname.replace(/^\[|\]$/g, "");
+    if (hostname.toLowerCase() === "localhost") {
+        throw new BlockedError("localhost is not a public website target");
+    }
+    const addresses = isIP(hostname) ? [hostname] : await resolve(hostname);
+    if (addresses.length === 0 || addresses.some((address) => !isPublicAddress(address))) {
+        throw new BlockedError("target resolved to a non-public address");
+    }
+    return addresses;
+}
+async function defaultResolver(hostname) {
+    const entries = await dnsLookup(hostname, { all: true, verbatim: true });
+    return entries.map((entry) => entry.address);
+}
+function requestWithPinnedLookup(url, addresses, signal) {
+    return new Promise((resolve, reject) => {
+        const transport = url.protocol === "https:" ? https : http;
+        const request = transport.request(url, {
+            method: "GET",
+            family: isIP(addresses[0]) || 4,
+            headers: { "user-agent": USER_AGENT, accept: "text/html,application/xhtml+xml,text/plain;q=0.8,*/*;q=0.1" },
+            lookup: (_hostname, _options, callback) => callback(null, addresses[0], isIP(addresses[0]) || 4)
+        }, (incoming) => {
+            const chunks = [];
+            let total = 0;
+            incoming.on("data", (chunk) => {
+                total += chunk.length;
+                if (total > MAX_HTML_BYTES) {
+                    request.destroy(new Error("response exceeded the maximum allowed size"));
+                    return;
+                }
+                chunks.push(chunk);
+            });
+            incoming.on("error", reject);
+            incoming.on("end", () => {
+                const headers = new Headers();
+                for (const [name, value] of Object.entries(incoming.headers)) {
+                    if (value !== undefined) {
+                        headers.set(name, Array.isArray(value) ? value.join(", ") : value);
+                    }
+                }
+                const status = incoming.statusCode ?? 500;
+                resolve(new Response([204, 205, 304].includes(status) ? null : Buffer.concat(chunks), { status, headers }));
+            });
+        });
+        const abort = () => request.destroy(new TimeoutError("request timed out"));
+        signal.addEventListener("abort", abort, { once: true });
+        request.on("close", () => signal.removeEventListener("abort", abort));
+        request.on("error", reject);
+        request.end();
+    });
+}
+async function readLimitedText(response) {
+    if (!response.body)
+        return "";
+    const reader = response.body.getReader();
+    const chunks = [];
+    let length = 0;
+    try {
+        for (;;) {
+            const { value, done } = await reader.read();
+            if (done)
+                break;
+            length += value.length;
+            if (length > MAX_HTML_BYTES)
+                throw new Error("response exceeded the maximum allowed size");
+            chunks.push(value);
+        }
+        return new TextDecoder().decode(Buffer.concat(chunks));
+    }
+    finally {
+        await reader.cancel();
+    }
+}
+function isRedirect(response) {
+    return response.status >= 300 && response.status < 400;
+}
+function contactLinks(html, baseUrl, origin, maxCount) {
+    const $ = load(html);
+    const candidates = [];
+    const signal = /contact|contact-us|iletisim|iletişim|kontakt|impressum|about|about-us|über|uber/i;
+    for (const anchor of $("a[href]").toArray()) {
+        const href = $(anchor).attr("href")?.trim();
+        if (!href || /^(mailto:|tel:|javascript:)/i.test(href)) {
+            continue;
+        }
+        const label = `${$(anchor).text()} ${$(anchor).attr("rel") ?? ""} ${href}`;
+        if (!signal.test(label)) {
+            continue;
+        }
+        try {
+            const candidate = new URL(href, baseUrl);
+            candidate.hash = "";
+            if (candidate.origin === origin && /^https?:$/.test(candidate.protocol) && !candidate.username && !candidate.password) {
+                candidates.push(candidate.toString());
+            }
+        }
+        catch {
+            continue;
+        }
+    }
+    return Array.from(new Set(candidates)).slice(0, maxCount);
+}
+function normalizeEmail(value) {
+    if (typeof value !== "string") {
+        return undefined;
+    }
+    const email = value.trim().replace(/^mailto:/i, "").split("?")[0]?.toLowerCase();
+    return email && /^[a-z0-9._%+-]+@[a-z0-9.-]+\.[a-z]{2,}$/i.test(email) ? email : undefined;
+}
+function normalizePhone(value) {
+    if (typeof value !== "string") {
+        return undefined;
+    }
+    const phone = value.trim().replace(/^tel:/i, "").replace(/[^\d+]/g, "");
+    return phone.replace(/\D/g, "").length >= 7 ? phone : undefined;
+}
+function socialProfile(value) {
+    if (typeof value !== "string") {
+        return undefined;
+    }
+    try {
+        const url = new URL(value);
+        const host = url.hostname.replace(/^www\./i, "").toLowerCase();
+        return ["facebook.com", "instagram.com", "linkedin.com", "x.com", "twitter.com", "tiktok.com", "youtube.com"].some((name) => host === name || host.endsWith(`.${name}`)) ? url.toString() : undefined;
+    }
+    catch {
+        return undefined;
+    }
+}
+function extractSchemaContact(html) {
+    const $ = load(html);
+    const emails = [];
+    const phones = [];
+    const socials = [];
+    const visit = (value) => {
+        if (Array.isArray(value)) {
+            value.forEach(visit);
+            return;
+        }
+        if (!value || typeof value !== "object") {
+            return;
+        }
+        const record = value;
+        const email = normalizeEmail(record.email);
+        const phone = normalizePhone(record.telephone);
+        if (email)
+            emails.push(email);
+        if (phone)
+            phones.push(phone);
+        const sameAs = Array.isArray(record.sameAs) ? record.sameAs : [record.sameAs];
+        sameAs.map(socialProfile).filter((url) => Boolean(url)).forEach((url) => socials.push(url));
+        visit(record.contactPoint);
+        visit(record["@graph"]);
+    };
+    for (const script of $("script[type='application/ld+json']").toArray()) {
+        try {
+            visit(JSON.parse($(script).text()));
+        }
+        catch {
+            continue;
+        }
+    }
+    return { publicEmail: emails[0], publicPhone: phones[0], socialProfiles: Array.from(new Set(socials)) };
+}
+function mergeContacts(entries) {
+    const populated = entries.filter(({ contact }) => contact.publicEmail || contact.publicPhone || contact.whatsappUrl || contact.contactPageUrl || contact.socialProfiles.length > 0);
+    if (populated.length === 0) {
+        return undefined;
+    }
+    const first = (key) => populated.map(({ contact }) => contact[key]).find(Boolean);
+    const socialProfiles = Array.from(new Set(populated.flatMap(({ contact }) => contact.socialProfiles)));
+    const channelCount = [first("publicEmail"), first("publicPhone"), first("whatsappUrl"), first("contactPageUrl"), ...socialProfiles].filter(Boolean).length;
+    const publicEmail = first("publicEmail");
+    const publicPhone = first("publicPhone");
+    return {
+        publicEmail,
+        publicPhone,
+        whatsappUrl: first("whatsappUrl"),
+        contactPageUrl: first("contactPageUrl"),
+        socialProfiles,
+        contactConfidence: publicEmail && publicPhone || channelCount >= 3 ? "High" : "Medium",
+        contactSource: populated.map(({ pageUrl }) => pageUrl).join(", ")
+    };
+}
+function combinePageContact(html, pageUrl) {
+    const contact = extractPublicContact(html, pageUrl);
+    const schema = extractSchemaContact(html);
+    return {
+        ...contact,
+        publicEmail: contact.publicEmail ?? schema.publicEmail,
+        publicPhone: contact.publicPhone ?? schema.publicPhone,
+        socialProfiles: Array.from(new Set([...contact.socialProfiles, ...schema.socialProfiles]))
+    };
+}
+export async function enrichWebsite(url, options = {}) {
+    const startedAt = Date.now();
+    const timeoutMs = options.timeoutMs ?? DEFAULT_TIMEOUT_MS;
+    const maxPages = Math.max(1, Math.min(options.maxPages ?? 3, 3));
+    const resolve = options.resolve ?? defaultResolver;
+    const robotsRules = new Map();
+    const warnings = [];
+    const sourceUrls = [];
+    let pagesFetched = 0;
+    const finish = (status, values = {}) => ({
+        status,
+        ...values,
+        sourceUrls,
+        pagesFetched,
+        ...(warnings.length ? { warnings } : {}),
+        durationMs: Date.now() - startedAt
+    });
+    const remainingMs = () => {
+        const remaining = timeoutMs - (Date.now() - startedAt);
+        if (remaining <= 0)
+            throw new TimeoutError("operation timed out");
+        return remaining;
+    };
+    const request = async (target) => {
+        const controller = new AbortController();
+        const requestTimeout = remainingMs();
+        const timer = setTimeout(() => controller.abort(), requestTimeout);
+        let raceTimer;
+        try {
+            const operation = (async () => {
+                const addresses = await resolvePublicAddresses(target.hostname, resolve);
+                if (controller.signal.aborted)
+                    throw new TimeoutError("request timed out");
+                const response = options.fetch
+                    ? await options.fetch(target.toString(), { redirect: "manual", signal: controller.signal, headers: { "user-agent": USER_AGENT } })
+                    : await requestWithPinnedLookup(target, addresses, controller.signal);
+                const body = await readLimitedText(response);
+                return new Response([204, 205, 304].includes(response.status) ? null : body, { status: response.status, headers: response.headers });
+            })();
+            return await Promise.race([
+                operation,
+                new Promise((_resolve, reject) => {
+                    raceTimer = setTimeout(() => reject(new TimeoutError("request timed out")), requestTimeout);
+                })
+            ]);
+        }
+        finally {
+            clearTimeout(timer);
+            if (raceTimer)
+                clearTimeout(raceTimer);
+        }
+    };
+    const allowedByRobots = async (target) => {
+        const cacheKey = target.origin;
+        const cached = robotsRules.get(cacheKey);
+        if (cached) {
+            if (cached.isAllowed(target.toString(), USER_AGENT) === false)
+                throw new BlockedError("robots.txt disallows this URL");
+            return;
+        }
+        let robotsUrl = new URL("/robots.txt", target.origin);
+        let response;
+        try {
+            response = await request(robotsUrl);
+            for (let count = 0; isRedirect(response) && count < MAX_REDIRECTS; count++) {
+                const location = response.headers.get("location");
+                if (!location)
+                    break;
+                const next = new URL(location, robotsUrl);
+                if (!/^https?:$/.test(next.protocol) || next.username || next.password || !sameBusinessHost(target.hostname, next.hostname))
+                    throw new BlockedError("robots redirect left the business website");
+                robotsUrl = next;
+                response = await request(robotsUrl);
+            }
+        }
+        catch (error) {
+            if (error instanceof BlockedError)
+                throw error;
+            if (error instanceof TimeoutError)
+                throw error;
+            throw new BlockedError("robots.txt could not be reached");
+        }
+        if (response.status === 404) {
+            robotsRules.set(cacheKey, robotsParser(new URL("/robots.txt", target.origin).toString(), ""));
+            return;
+        }
+        if (response.status === 403 || response.status < 200 || response.status >= 300) {
+            throw new BlockedError("robots.txt did not permit crawling");
+        }
+        const policy = robotsParser(new URL("/robots.txt", target.origin).toString(), await readLimitedText(response));
+        robotsRules.set(cacheKey, policy);
+        if (policy.isAllowed(target.toString(), USER_AGENT) === false) {
+            throw new BlockedError("robots.txt disallows this URL");
+        }
+    };
+    const fetchPage = async (requestedUrl, homepageHost) => {
+        let current = new URL(requestedUrl);
+        for (let redirects = 0; redirects <= MAX_REDIRECTS; redirects += 1) {
+            if (!/^https?:$/.test(current.protocol) || current.username || current.password || !sameBusinessHost(homepageHost, current.hostname)) {
+                throw new BlockedError("redirect left the business website");
+            }
+            await allowedByRobots(current);
+            const response = await request(current);
+            if (isRedirect(response)) {
+                const location = response.headers.get("location");
+                if (!location)
+                    throw new Error("redirect response had no location");
+                current = new URL(location, current);
+                continue;
+            }
+            if (response.status === 403)
+                throw new BlockedError("website denied the request");
+            if (response.status === 429)
+                throw new Error("website rate limited the request");
+            if (!response.ok)
+                throw new Error(`website returned HTTP ${response.status}`);
+            const contentType = response.headers.get("content-type")?.toLowerCase();
+            if (contentType && !contentType.includes("text/html") && !contentType.includes("application/xhtml+xml")) {
+                throw new Error("website response was not HTML");
+            }
+            const html = await readLimitedText(response);
+            pagesFetched += 1;
+            if (!sourceUrls.includes(current.toString()))
+                sourceUrls.push(current.toString());
+            return { url: requestedUrl, finalUrl: current.toString(), statusCode: response.status, headers: normalizeHeaders(response.headers), html };
+        }
+        throw new Error(`exceeded redirect limit of ${MAX_REDIRECTS}`);
+    };
+    try {
+        const initial = new URL(url);
+        if (!/^https?:$/.test(initial.protocol) || initial.username || initial.password) {
+            throw new BlockedError("only public HTTP(S) URLs are allowed");
+        }
+        const homepage = await fetchPage(initial.toString(), initial.hostname);
+        const contactEntries = [{ contact: combinePageContact(homepage.html, homepage.finalUrl), pageUrl: homepage.finalUrl }];
+        const followUps = contactLinks(homepage.html, homepage.finalUrl, new URL(homepage.finalUrl).origin, maxPages - 1);
+        for (const followUp of followUps) {
+            try {
+                const page = await fetchPage(followUp, initial.hostname);
+                contactEntries.push({ contact: combinePageContact(page.html, page.finalUrl), pageUrl: page.finalUrl });
+            }
+            catch (error) {
+                warnings.push(`${followUp}: ${error instanceof Error ? error.message : "could not read page"}`);
+                break;
+            }
+        }
+        return finish("success", { snapshot: homepage, contact: mergeContacts(contactEntries) });
+    }
+    catch (error) {
+        if (error instanceof BlockedError)
+            return finish("blocked", { error: error.message });
+        return finish("failed", { error: error instanceof Error ? error.message : "website enrichment failed" });
+    }
+}
+//# sourceMappingURL=website-enrichment.js.map