gasp-standard-mcp 1.0.0

package/build/parser.js

@@ -0,0 +1,256 @@
+/** Map of file basenames to department display names */
+const DEPARTMENT_NAMES = {
+    core: "Core",
+    sales: "Sales",
+    marketing: "Marketing",
+    "customer-success": "Customer Success",
+    support: "Support",
+    product: "Product",
+    engineering: "Engineering",
+    finance: "Finance",
+    onboarding: "Onboarding",
+    people: "People",
+    revops: "RevOps",
+    partnerships: "Partnerships",
+    "professional-services": "Professional Services",
+};
+/** Known abbreviation → full name mapping */
+const ABBREVIATION_MAP = {
+    MRR: ["Monthly Recurring Revenue"],
+    ARR: ["Annual Recurring Revenue"],
+    NRR: ["Net Revenue Retention", "NDR", "Net Dollar Retention"],
+    NDR: ["Net Dollar Retention", "NRR", "Net Revenue Retention"],
+    GRR: ["Gross Revenue Retention"],
+    CAC: ["Customer Acquisition Cost"],
+    LTV: ["Customer Lifetime Value", "CLV", "CLTV"],
+    CLV: ["Customer Lifetime Value", "LTV", "CLTV"],
+    CLTV: ["Customer Lifetime Value", "LTV", "CLV"],
+    NPS: ["Net Promoter Score"],
+    CSAT: ["Customer Satisfaction Score"],
+    CES: ["Customer Effort Score"],
+    ARPA: ["Average Revenue Per Account"],
+    ACV: ["Annual Contract Value"],
+    TTFV: ["Time to First Value"],
+    FRT: ["First Response Time"],
+    MTTR: ["Mean Time to Resolve", "Mean Time to Recovery"],
+    DAU: ["Daily Active Users"],
+    MAU: ["Monthly Active Users"],
+    WAU: ["Weekly Active Users"],
+    MQL: ["Marketing Qualified Lead"],
+    SQL: ["Sales Qualified Lead"],
+    DSO: ["Days Sales Outstanding"],
+    EBITDA: ["Earnings Before Interest Taxes Depreciation and Amortization"],
+    COGS: ["Cost of Goods Sold"],
+    DORA: ["DevOps Research and Assessment"],
+};
+/**
+ * Extract abbreviation from metric name heading.
+ * e.g. "MRR (Monthly Recurring Revenue)" → ["MRR"]
+ * e.g. "Net Revenue Retention (NRR)" → ["NRR", "NDR"]
+ */
+function extractAliases(name) {
+    const aliases = [];
+    // Match abbreviation at start: "MRR (Monthly Recurring Revenue)"
+    const leadingMatch = name.match(/^([A-Z]{2,})\s/);
+    if (leadingMatch) {
+        aliases.push(leadingMatch[1]);
+    }
+    // Match abbreviation in parens: "Net Revenue Retention (NRR)"
+    const parenMatch = name.match(/\(([A-Z]{2,})\)/);
+    if (parenMatch) {
+        aliases.push(parenMatch[1]);
+    }
+    // Add known aliases for any found abbreviation
+    for (const alias of [...aliases]) {
+        const known = ABBREVIATION_MAP[alias];
+        if (known) {
+            for (const k of known) {
+                if (k.length <= 5 && !aliases.includes(k)) {
+                    aliases.push(k);
+                }
+            }
+        }
+    }
+    return aliases;
+}
+/** Slugify a heading for anchor links */
+function slugify(text) {
+    return text
+        .toLowerCase()
+        .replace(/[^a-z0-9\s-]/g, "")
+        .replace(/\s+/g, "-")
+        .replace(/-+/g, "-")
+        .trim();
+}
+/**
+ * Parse a single metric markdown file into an array of Metric objects.
+ * Splits on ### headings and extracts fields.
+ */
+export function parseMetricsFile(content, filename) {
+    const dept = DEPARTMENT_NAMES[filename.replace(".md", "")] || filename.replace(".md", "");
+    const metrics = [];
+    let currentCategory = "";
+    // Split into lines and walk through
+    const lines = content.split("\n");
+    let i = 0;
+    while (i < lines.length) {
+        const line = lines[i];
+        // Track ## section headings as categories
+        if (line.startsWith("## ") && !line.startsWith("### ")) {
+            currentCategory = line.replace("## ", "").trim();
+            i++;
+            continue;
+        }
+        // Found a ### metric heading
+        if (line.startsWith("### ")) {
+            const name = line.replace("### ", "").trim();
+            const slug = slugify(name);
+            const aliases = extractAliases(name);
+            // Collect all lines until next ### or ## or end
+            const sectionLines = [];
+            i++;
+            while (i < lines.length &&
+                !lines[i].startsWith("### ") &&
+                !(lines[i].startsWith("## ") && !lines[i].startsWith("### "))) {
+                sectionLines.push(lines[i]);
+                i++;
+            }
+            const rawMarkdown = sectionLines.join("\n").trim();
+            const metric = parseSectionFields(name, slug, rawMarkdown, dept, currentCategory, aliases);
+            metrics.push(metric);
+            continue;
+        }
+        i++;
+    }
+    return metrics;
+}
+/** Extract structured fields from a metric section's markdown body */
+function parseSectionFields(name, slug, raw, department, category, aliases) {
+    const definition = extractField(raw, "Definition");
+    const insight = extractField(raw, "What it tells you");
+    const benchmarks = extractBenchmarks(raw);
+    const formulas = extractFormulas(raw);
+    const commonMistakes = extractListField(raw, "Common mistakes");
+    const sources = extractSources(raw);
+    return {
+        name,
+        slug,
+        definition: definition || "",
+        formulas,
+        benchmarks,
+        insight,
+        commonMistakes,
+        sources,
+        department,
+        category,
+        aliases,
+        rawMarkdown: `### ${name}\n\n${raw}`,
+    };
+}
+/** Extract a **FieldName:** value from markdown */
+function extractField(text, fieldName) {
+    const fence = "\\`\\`\\`";
+    const pattern = new RegExp("\\*\\*" + fieldName + ":\\*\\*\\s*(.+?)(?=\\n\\n|\\n\\*\\*|\\n" + fence + "|$)", "s");
+    const match = text.match(pattern);
+    return match ? match[1].trim() : null;
+}
+/** Extract all code block formulas */
+function extractFormulas(text) {
+    const formulas = [];
+    const fence = "```";
+    const codeBlockRegex = new RegExp(fence + "\\n?([\\s\\S]*?)" + fence, "g");
+    let match;
+    while ((match = codeBlockRegex.exec(text)) !== null) {
+        const formula = match[1].trim();
+        if (formula) {
+            formulas.push(formula);
+        }
+    }
+    return formulas;
+}
+/** Extract benchmarks section — everything after **Benchmarks until next field */
+function extractBenchmarks(text) {
+    const pattern = /\*\*Benchmarks[^*]*\*\*\s*([\s\S]*?)(?=\n\*\*[A-Z]|\n##|\n---|\n\*\*Sources|$)/;
+    const match = text.match(pattern);
+    if (!match)
+        return null;
+    return match[1].trim() || null;
+}
+/** Extract a bulleted list after **FieldName:** */
+function extractListField(text, fieldName) {
+    const pattern = new RegExp(`\\*\\*${fieldName}:\\*\\*\\s*\\n((?:- .+\\n?)+)`, "m");
+    const match = text.match(pattern);
+    if (!match)
+        return null;
+    return match[1]
+        .split("\n")
+        .filter((l) => l.startsWith("- "))
+        .map((l) => l.replace(/^- /, "").trim());
+}
+/** Extract source URLs */
+function extractSources(text) {
+    const sources = [];
+    const pattern = /\*\*Sources:\*\*\s*\n([\s\S]*?)(?=\n\*\*[A-Z]|\n##|\n---|\n$|$)/;
+    const match = text.match(pattern);
+    if (!match)
+        return sources;
+    const linkRegex = /\[([^\]]+)\]\(([^)]+)\)/g;
+    let linkMatch;
+    while ((linkMatch = linkRegex.exec(match[1])) !== null) {
+        sources.push(`${linkMatch[1]}: ${linkMatch[2]}`);
+    }
+    return sources;
+}
+/**
+ * Parse definitions.md into Definition objects.
+ * Each ### heading is a term.
+ */
+export function parseDefinitions(content) {
+    const definitions = [];
+    const lines = content.split("\n");
+    let currentSection = "";
+    let i = 0;
+    while (i < lines.length) {
+        const line = lines[i];
+        if (line.startsWith("## ") && !line.startsWith("### ")) {
+            currentSection = line.replace("## ", "").trim();
+            i++;
+            continue;
+        }
+        if (line.startsWith("### ")) {
+            const term = line.replace("### ", "").trim();
+            const slug = slugify(term);
+            const sectionLines = [];
+            i++;
+            while (i < lines.length &&
+                !lines[i].startsWith("### ") &&
+                !(lines[i].startsWith("## ") && !lines[i].startsWith("### "))) {
+                sectionLines.push(lines[i]);
+                i++;
+            }
+            const raw = sectionLines.join("\n").trim();
+            // First non-empty line after heading is the definition
+            const defLine = raw.split("\n").find((l) => l.trim() && !l.startsWith("---"));
+            const definition = defLine?.trim() || "";
+            // Extract canonical link
+            const linkMatch = raw.match(/\*\*Canonical definition:\*\*\s*\[([^\]]+)\]\(([^)]+)\)/);
+            const canonicalLink = linkMatch ? linkMatch[2] : null;
+            // Extract "Used in" departments
+            const usedInMatch = raw.match(/\*\*Used in:\*\*\s*(.+)/);
+            const usedIn = usedInMatch
+                ? usedInMatch[1].split(",").map((s) => s.trim())
+                : [];
+            definitions.push({
+                term,
+                slug,
+                definition,
+                section: currentSection,
+                canonicalLink,
+                usedIn,
+            });
+            continue;
+        }
+        i++;
+    }
+    return definitions;
+}