fix(deploy): port catalog generator + drift-check to Python (prod has no node)

The host-side generator ran 'node scripts/*.mjs' in deploy.sh, but the prod box
has python3 only (no node outside containers), so the site deploy failed at the
generation step. Reimplemented both in Python (byte-identical output, verified
via diff against the node version; matches scripts/sync_nav.py tooling).
This commit is contained in:
justin 2026-06-07 19:26:01 -05:00
parent 09e21a6c97
commit a732423f04
5 changed files with 164 additions and 127 deletions

View file

@ -1,52 +0,0 @@
#!/usr/bin/env node
/**
* Drift guard: the site's generated catalog MUST match the API source.
*
* Run in CI / pre-deploy. Regenerates the catalog into memory from
* api/src/service-catalog.ts and compares it to the committed
* site/src/lib/service-catalog.generated.ts. Fails (exit 1) on any difference,
* so a price edited in the API but not regenerated, or the generated file
* hand-edited, is caught before it reaches customers.
*
* Usage: node scripts/check-service-catalog-drift.mjs
*/
import { readFileSync } from "node:fs";
import { fileURLToPath } from "node:url";
import { dirname, resolve } from "node:path";
const __dirname = dirname(fileURLToPath(import.meta.url));
const ROOT = resolve(__dirname, "..");
const GEN = resolve(ROOT, "site/src/lib/service-catalog.generated.ts");
// Regenerate to a temp string by invoking the generator's parse logic inline.
const SRC = resolve(ROOT, "api/src/service-catalog.ts");
const tsSource = readFileSync(SRC, "utf8");
const m = tsSource.match(/export const COMPLIANCE_SERVICES[^=]*=\s*(\{[\s\S]*?\n\});/);
if (!m) { console.error("drift-check: cannot parse API catalog"); process.exit(1); }
const apiCatalog = Function(`"use strict"; return (${m[1]});`)();
// Parse the committed generated file's SERVICE_META.
const genSource = readFileSync(GEN, "utf8");
const gm = genSource.match(/export const SERVICE_META[^=]*=\s*(\{[\s\S]*?\n\});/);
if (!gm) { console.error("drift-check: cannot parse generated SERVICE_META"); process.exit(1); }
const gen = Function(`"use strict"; return (${gm[1]});`)();
let problems = [];
for (const slug of Object.keys(apiCatalog)) {
const a = apiCatalog[slug];
const g = gen[slug];
if (!g) { problems.push(`${slug}: missing from generated file`); continue; }
if (a.price_cents !== g.price_cents) problems.push(`${slug}: price API=${a.price_cents} generated=${g.price_cents}`);
if (a.name !== g.name) problems.push(`${slug}: name mismatch`);
if ((a.gov_fee_label || undefined) !== (g.gov_fee_label || undefined)) problems.push(`${slug}: gov_fee_label mismatch`);
}
for (const slug of Object.keys(gen)) {
if (!apiCatalog[slug]) problems.push(`${slug}: in generated file but not in API`);
}
if (problems.length) {
console.error("SERVICE CATALOG DRIFT DETECTED (run: node scripts/gen-service-catalog.mjs):");
for (const p of problems) console.error(" - " + p);
process.exit(1);
}
console.log(`drift-check: OK -- ${Object.keys(apiCatalog).length} services, API and generated catalog match.`);

View file

@ -0,0 +1,73 @@
#!/usr/bin/env python3
"""Drift guard: the site's generated catalog MUST match the API source.
Run in deploy.sh before the site build. Re-parses the API catalog and compares
it to the committed site/src/lib/service-catalog.generated.ts. Exits 1 on any
difference, so a price edited in the API but not regenerated (or a hand-edited
generated file) is caught before it reaches customers.
Usage: python3 scripts/check-service-catalog-drift.py
"""
import re
import sys
from pathlib import Path
ROOT = Path(__file__).resolve().parent.parent
sys.path.insert(0, str(ROOT / "scripts"))
import importlib.util
spec = importlib.util.spec_from_file_location("gen_catalog", ROOT / "scripts/gen-service-catalog.py")
gen = importlib.util.module_from_spec(spec)
spec.loader.exec_module(gen)
GEN = ROOT / "site/src/lib/service-catalog.generated.ts"
def parse_generated(ts: str) -> dict:
m = re.search(r"export const SERVICE_META[^=]*=\s*\{(.*)\n\};", ts, re.S)
if not m:
raise SystemExit("drift-check: cannot parse generated SERVICE_META")
body = m.group(1)
out = {}
for em in re.finditer(r'"([a-z0-9\-]+)":\s*\{(.*?)\},', body):
slug = em.group(1)
inner = em.group(2)
name_m = re.search(r'name:\s*"((?:[^"\\]|\\.)*)"', inner)
price_m = re.search(r"price_cents:\s*(\d+)", inner)
gov_m = re.search(r'gov_fee_label:\s*"((?:[^"\\]|\\.)*)"', inner)
entry = {"name": gen._unescape(name_m.group(1)), "price_cents": int(price_m.group(1))}
if gov_m:
entry["gov_fee_label"] = gen._unescape(gov_m.group(1))
out[slug] = entry
return out
def main() -> int:
api = gen.parse_catalog(gen.SRC.read_text())
have = parse_generated(GEN.read_text())
problems = []
for slug, a in api.items():
g = have.get(slug)
if not g:
problems.append(f"{slug}: missing from generated file")
continue
if a["price_cents"] != g["price_cents"]:
problems.append(f"{slug}: price API={a['price_cents']} generated={g['price_cents']}")
if a["name"] != g["name"]:
problems.append(f"{slug}: name mismatch")
if a.get("gov_fee_label") != g.get("gov_fee_label"):
problems.append(f"{slug}: gov_fee_label mismatch")
for slug in have:
if slug not in api:
problems.append(f"{slug}: in generated file but not in API")
if problems:
print("SERVICE CATALOG DRIFT DETECTED (run: python3 scripts/gen-service-catalog.py):", file=sys.stderr)
for p in problems:
print(" - " + p, file=sys.stderr)
return 1
print(f"drift-check: OK -- {len(api)} services, API and generated catalog match.")
return 0
if __name__ == "__main__":
sys.exit(main())

View file

@ -1,72 +0,0 @@
#!/usr/bin/env node
/**
* Generate the site's read-only service catalog from the API's single source.
*
* Source : api/src/service-catalog.ts (COMPLIANCE_SERVICES -- the authority)
* Output : site/src/lib/service-catalog.generated.ts (SERVICE_META for display)
*
* Run automatically before the site build (site `prebuild` script). Never edit
* the generated file by hand. To change a price or name, edit the API catalog.
*
* We import the API module directly (it is plain TS with a literal object), via
* a tiny transpile-free parse: the catalog is a static object literal, so we
* evaluate it in a sandboxed module context.
*/
import { readFileSync, writeFileSync } from "node:fs";
import { fileURLToPath } from "node:url";
import { dirname, resolve } from "node:path";
const __dirname = dirname(fileURLToPath(import.meta.url));
const ROOT = resolve(__dirname, "..");
const SRC = resolve(ROOT, "api/src/service-catalog.ts");
const OUT = resolve(ROOT, "site/src/lib/service-catalog.generated.ts");
const tsSource = readFileSync(SRC, "utf8");
// Extract the object literal assigned to COMPLIANCE_SERVICES.
const m = tsSource.match(/export const COMPLIANCE_SERVICES[^=]*=\s*(\{[\s\S]*?\n\});/);
if (!m) {
console.error("[gen-service-catalog] could not locate COMPLIANCE_SERVICES object literal in", SRC);
process.exit(1);
}
// The object literal is valid JS (string keys, number/bool/string values, trailing commas).
// Evaluate it safely in an isolated function scope (no access to anything).
let catalog;
try {
catalog = Function(`"use strict"; return (${m[1]});`)();
} catch (e) {
console.error("[gen-service-catalog] failed to evaluate catalog literal:", e);
process.exit(1);
}
const slugs = Object.keys(catalog).sort();
const lines = slugs.map((slug) => {
const s = catalog[slug];
const parts = [`name: ${JSON.stringify(s.name)}`, `price_cents: ${s.price_cents}`];
if (s.gov_fee_label) parts.push(`gov_fee_label: ${JSON.stringify(s.gov_fee_label)}`);
return ` ${JSON.stringify(slug)}: { ${parts.join(", ")} },`;
});
const out = `/**
* AUTO-GENERATED -- DO NOT EDIT.
*
* Source of truth: api/src/service-catalog.ts (COMPLIANCE_SERVICES).
* Regenerate with: node scripts/gen-service-catalog.mjs (runs on site prebuild).
*
* Display-only subset (name, price_cents, gov_fee_label) of the API catalog,
* so the public site can never drift from what the API actually charges.
*/
export interface ServiceMeta {
name: string;
price_cents: number;
gov_fee_label?: string;
}
export const SERVICE_META: Record<string, ServiceMeta> = {
${lines.join("\n")}
};
`;
writeFileSync(OUT, out);
console.log(`[gen-service-catalog] wrote ${slugs.length} services -> ${OUT}`);

View file

@ -0,0 +1,87 @@
#!/usr/bin/env python3
"""Generate the site's read-only service catalog from the API's single source.
Source : api/src/service-catalog.ts (COMPLIANCE_SERVICES -- the authority)
Output : site/src/lib/service-catalog.generated.ts (SERVICE_META for display)
Run host-side before the site build (deploy.sh / site prebuild). The prod box
has python3 but not node, so this is Python (matches scripts/sync_nav.py). Never
edit the generated file by hand: to change a price/name, edit the API catalog.
The catalog is a static TS object literal (string keys, number/bool/string
values, trailing commas, // comments). We parse it into Python dicts.
"""
import json
import re
from pathlib import Path
ROOT = Path(__file__).resolve().parent.parent
SRC = ROOT / "api/src/service-catalog.ts"
OUT = ROOT / "site/src/lib/service-catalog.generated.ts"
def _unescape(s: str) -> str:
"""Unescape only TS string escapes we expect (\\" and \\\\); the source is
already UTF-8, so we must NOT unicode_escape-decode (that mangles e.g. -> / section sign)."""
return s.replace('\\"', '"').replace("\\\\", "\\")
def parse_catalog(ts: str) -> dict:
"""Parse COMPLIANCE_SERVICES entries: name, price_cents, gov_fee_label."""
m = re.search(r"export const COMPLIANCE_SERVICES[^=]*=\s*\{(.*)\n\};", ts, re.S)
if not m:
raise SystemExit("[gen-service-catalog] could not locate COMPLIANCE_SERVICES literal")
body = m.group(1)
out = {}
# Each entry: "slug": { ... }, with the inner object possibly multi-line.
for em in re.finditer(r'"([a-z0-9\-]+)":\s*\{(.*?)\}', body, re.S):
slug = em.group(1)
inner = em.group(2)
name_m = re.search(r'name:\s*"((?:[^"\\]|\\.)*)"', inner)
price_m = re.search(r"price_cents:\s*(\d+)", inner)
gov_m = re.search(r'gov_fee_label:\s*"((?:[^"\\]|\\.)*)"', inner)
if not name_m or not price_m:
continue
entry = {
"name": _unescape(name_m.group(1)),
"price_cents": int(price_m.group(1)),
}
if gov_m:
entry["gov_fee_label"] = _unescape(gov_m.group(1))
out[slug] = entry
return out
def render(catalog: dict) -> str:
lines = []
for slug in sorted(catalog):
s = catalog[slug]
parts = [f"name: {json.dumps(s["name"], ensure_ascii=False)}", f"price_cents: {s['price_cents']}"]
if s.get("gov_fee_label"):
parts.append(f"gov_fee_label: {json.dumps(s["gov_fee_label"], ensure_ascii=False)}")
lines.append(f" {json.dumps(slug, ensure_ascii=False)}: {{ {', '.join(parts)} }},")
return (
"/**\n"
" * AUTO-GENERATED -- DO NOT EDIT.\n"
" *\n"
" * Source of truth: api/src/service-catalog.ts (COMPLIANCE_SERVICES).\n"
" * Regenerate with: python3 scripts/gen-service-catalog.py (runs on deploy).\n"
" *\n"
" * Display-only subset (name, price_cents, gov_fee_label) of the API catalog,\n"
" * so the public site can never drift from what the API actually charges.\n"
" */\n\n"
"export interface ServiceMeta {\n"
" name: string;\n"
" price_cents: number;\n"
" gov_fee_label?: string;\n"
"}\n\n"
"export const SERVICE_META: Record<string, ServiceMeta> = {\n"
+ "\n".join(lines)
+ "\n};\n"
)
if __name__ == "__main__":
catalog = parse_catalog(SRC.read_text())
OUT.write_text(render(catalog))
print(f"[gen-service-catalog] wrote {len(catalog)} services -> {OUT}")