xtrawl 0.1.1 → 0.1.3
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/DOCUMENTATION.md +51 -0
- package/README.md +34 -6
- package/dist/auth/bootstrap.d.ts +1 -1
- package/dist/auth/bootstrap.d.ts.map +1 -1
- package/dist/auth/bootstrap.js +27 -14
- package/dist/auth/bootstrap.js.map +1 -1
- package/dist/cli/main.d.ts.map +1 -1
- package/dist/cli/main.js +13 -12
- package/dist/cli/main.js.map +1 -1
- package/dist/cli/parser.d.ts +2 -14
- package/dist/cli/parser.d.ts.map +1 -1
- package/dist/cli/parser.js +26 -30
- package/dist/cli/parser.js.map +1 -1
- package/dist/cli/types.d.ts +17 -0
- package/dist/cli/types.d.ts.map +1 -0
- package/dist/cli/types.js +2 -0
- package/dist/cli/types.js.map +1 -0
- package/dist/client/accounts.d.ts.map +1 -1
- package/dist/client/accounts.js +13 -6
- package/dist/client/accounts.js.map +1 -1
- package/dist/client/client.d.ts +16 -3
- package/dist/client/client.d.ts.map +1 -1
- package/dist/client/client.js +96 -38
- package/dist/client/client.js.map +1 -1
- package/dist/client/collectors.d.ts +2 -10
- package/dist/client/collectors.d.ts.map +1 -1
- package/dist/client/collectors.js +122 -57
- package/dist/client/collectors.js.map +1 -1
- package/dist/client/database.d.ts +6 -18
- package/dist/client/database.d.ts.map +1 -1
- package/dist/client/database.js +4 -21
- package/dist/client/database.js.map +1 -1
- package/dist/client/page-commit.d.ts +15 -0
- package/dist/client/page-commit.d.ts.map +1 -0
- package/dist/client/page-commit.js +19 -0
- package/dist/client/page-commit.js.map +1 -0
- package/dist/client/profiles.d.ts +1 -1
- package/dist/client/profiles.d.ts.map +1 -1
- package/dist/client/profiles.js +10 -3
- package/dist/client/profiles.js.map +1 -1
- package/dist/client/search.d.ts +5 -17
- package/dist/client/search.d.ts.map +1 -1
- package/dist/client/search.js +81 -26
- package/dist/client/search.js.map +1 -1
- package/dist/client/types.d.ts +40 -2
- package/dist/client/types.d.ts.map +1 -1
- package/dist/config/defaults.d.ts.map +1 -1
- package/dist/config/defaults.js +5 -2
- package/dist/config/defaults.js.map +1 -1
- package/dist/config/types.d.ts +5 -2
- package/dist/config/types.d.ts.map +1 -1
- package/dist/config/validation.d.ts +1 -1
- package/dist/config/validation.d.ts.map +1 -1
- package/dist/config/validation.js +27 -11
- package/dist/config/validation.js.map +1 -1
- package/dist/constants/accounts.d.ts +16 -0
- package/dist/constants/accounts.d.ts.map +1 -0
- package/dist/constants/accounts.js +16 -0
- package/dist/constants/accounts.js.map +1 -0
- package/dist/constants/cli.d.ts +10 -0
- package/dist/constants/cli.d.ts.map +1 -0
- package/dist/constants/cli.js +10 -0
- package/dist/constants/cli.js.map +1 -0
- package/dist/constants/collection.d.ts +9 -0
- package/dist/constants/collection.d.ts.map +1 -0
- package/dist/constants/collection.js +9 -0
- package/dist/constants/collection.js.map +1 -0
- package/dist/constants/config.d.ts +6 -0
- package/dist/constants/config.d.ts.map +1 -0
- package/dist/constants/config.js +6 -0
- package/dist/constants/config.js.map +1 -0
- package/dist/constants/errors.d.ts +9 -0
- package/dist/constants/errors.d.ts.map +1 -0
- package/dist/constants/errors.js +9 -0
- package/dist/constants/errors.js.map +1 -0
- package/dist/constants/output.d.ts +7 -0
- package/dist/constants/output.d.ts.map +1 -0
- package/dist/constants/output.js +7 -0
- package/dist/constants/output.js.map +1 -0
- package/dist/constants/pages.d.ts +5 -0
- package/dist/constants/pages.d.ts.map +1 -0
- package/dist/constants/pages.js +5 -0
- package/dist/constants/pages.js.map +1 -0
- package/dist/constants/requests.d.ts +18 -0
- package/dist/constants/requests.d.ts.map +1 -0
- package/dist/constants/requests.js +18 -0
- package/dist/constants/requests.js.map +1 -0
- package/dist/constants/runner.d.ts +8 -0
- package/dist/constants/runner.d.ts.map +1 -0
- package/dist/constants/runner.js +8 -0
- package/dist/constants/runner.js.map +1 -0
- package/dist/constants/runs.d.ts +8 -0
- package/dist/constants/runs.d.ts.map +1 -0
- package/dist/constants/runs.js +8 -0
- package/dist/constants/runs.js.map +1 -0
- package/dist/domain/accounts.d.ts +4 -2
- package/dist/domain/accounts.d.ts.map +1 -1
- package/dist/domain/collection.d.ts +27 -0
- package/dist/domain/collection.d.ts.map +1 -0
- package/dist/domain/collection.js +2 -0
- package/dist/domain/collection.js.map +1 -0
- package/dist/domain/diagnostics.d.ts +11 -0
- package/dist/domain/diagnostics.d.ts.map +1 -0
- package/dist/domain/diagnostics.js +2 -0
- package/dist/domain/diagnostics.js.map +1 -0
- package/dist/domain/errors.d.ts +2 -1
- package/dist/domain/errors.d.ts.map +1 -1
- package/dist/domain/errors.js +10 -9
- package/dist/domain/errors.js.map +1 -1
- package/dist/domain/http.d.ts +7 -1
- package/dist/domain/http.d.ts.map +1 -1
- package/dist/domain/manifest.d.ts +8 -0
- package/dist/domain/manifest.d.ts.map +1 -1
- package/dist/domain/output.d.ts +3 -0
- package/dist/domain/output.d.ts.map +1 -0
- package/dist/domain/output.js +2 -0
- package/dist/domain/output.js.map +1 -0
- package/dist/domain/pages.d.ts +20 -0
- package/dist/domain/pages.d.ts.map +1 -0
- package/dist/domain/pages.js +2 -0
- package/dist/domain/pages.js.map +1 -0
- package/dist/domain/pool.d.ts +22 -0
- package/dist/domain/pool.d.ts.map +1 -0
- package/dist/domain/pool.js +2 -0
- package/dist/domain/pool.js.map +1 -0
- package/dist/domain/records.d.ts +6 -1
- package/dist/domain/records.d.ts.map +1 -1
- package/dist/domain/requests.d.ts +19 -10
- package/dist/domain/requests.d.ts.map +1 -1
- package/dist/domain/runner.d.ts +30 -0
- package/dist/domain/runner.d.ts.map +1 -0
- package/dist/domain/runner.js +2 -0
- package/dist/domain/runner.js.map +1 -0
- package/dist/domain/runs.d.ts +12 -0
- package/dist/domain/runs.d.ts.map +1 -0
- package/dist/domain/runs.js +2 -0
- package/dist/domain/runs.js.map +1 -0
- package/dist/engine/api-engine.d.ts +7 -7
- package/dist/engine/api-engine.d.ts.map +1 -1
- package/dist/engine/api-engine.js +42 -19
- package/dist/engine/api-engine.js.map +1 -1
- package/dist/engine/extractors.d.ts +2 -8
- package/dist/engine/extractors.d.ts.map +1 -1
- package/dist/engine/extractors.js +30 -11
- package/dist/engine/extractors.js.map +1 -1
- package/dist/index.d.ts +4 -4
- package/dist/index.d.ts.map +1 -1
- package/dist/index.js.map +1 -1
- package/dist/manifest/destinations.d.ts +4 -0
- package/dist/manifest/destinations.d.ts.map +1 -0
- package/dist/manifest/destinations.js +47 -0
- package/dist/manifest/destinations.js.map +1 -0
- package/dist/manifest/model.d.ts +3 -2
- package/dist/manifest/model.d.ts.map +1 -1
- package/dist/manifest/model.js +6 -1
- package/dist/manifest/model.js.map +1 -1
- package/dist/manifest/provider.d.ts +10 -3
- package/dist/manifest/provider.d.ts.map +1 -1
- package/dist/manifest/provider.js +86 -34
- package/dist/manifest/provider.js.map +1 -1
- package/dist/manifest/scraper.d.ts +2 -6
- package/dist/manifest/scraper.d.ts.map +1 -1
- package/dist/manifest/scraper.js +46 -15
- package/dist/manifest/scraper.js.map +1 -1
- package/dist/output/csv-writer.d.ts.map +1 -1
- package/dist/output/csv-writer.js +32 -10
- package/dist/output/csv-writer.js.map +1 -1
- package/dist/output/json-writer.d.ts +1 -0
- package/dist/output/json-writer.d.ts.map +1 -1
- package/dist/output/json-writer.js +36 -11
- package/dist/output/json-writer.js.map +1 -1
- package/dist/output/lock.d.ts +2 -0
- package/dist/output/lock.d.ts.map +1 -0
- package/dist/output/lock.js +18 -0
- package/dist/output/lock.js.map +1 -0
- package/dist/output/writer.d.ts +2 -1
- package/dist/output/writer.d.ts.map +1 -1
- package/dist/output/writer.js +6 -3
- package/dist/output/writer.js.map +1 -1
- package/dist/pool/account-pool.d.ts +3 -12
- package/dist/pool/account-pool.d.ts.map +1 -1
- package/dist/pool/account-pool.js +54 -16
- package/dist/pool/account-pool.js.map +1 -1
- package/dist/pool/cooldown.d.ts +3 -7
- package/dist/pool/cooldown.d.ts.map +1 -1
- package/dist/pool/cooldown.js +10 -7
- package/dist/pool/cooldown.js.map +1 -1
- package/dist/pool/limiter.d.ts +3 -3
- package/dist/pool/limiter.d.ts.map +1 -1
- package/dist/pool/limiter.js +5 -6
- package/dist/pool/limiter.js.map +1 -1
- package/dist/query/builder.d.ts +1 -1
- package/dist/query/builder.d.ts.map +1 -1
- package/dist/query/builder.js +10 -4
- package/dist/query/builder.js.map +1 -1
- package/dist/query/collection-id.d.ts +3 -0
- package/dist/query/collection-id.d.ts.map +1 -0
- package/dist/query/collection-id.js +21 -0
- package/dist/query/collection-id.js.map +1 -0
- package/dist/query/normalize.d.ts +3 -32
- package/dist/query/normalize.d.ts.map +1 -1
- package/dist/query/normalize.js +10 -16
- package/dist/query/normalize.js.map +1 -1
- package/dist/query/targets.d.ts +2 -7
- package/dist/query/targets.d.ts.map +1 -1
- package/dist/query/targets.js.map +1 -1
- package/dist/query/types.d.ts +40 -0
- package/dist/query/types.d.ts.map +1 -0
- package/dist/query/types.js +2 -0
- package/dist/query/types.js.map +1 -0
- package/dist/runner/index.d.ts +3 -0
- package/dist/runner/index.d.ts.map +1 -1
- package/dist/runner/index.js.map +1 -1
- package/dist/runner/retry.d.ts +3 -6
- package/dist/runner/retry.d.ts.map +1 -1
- package/dist/runner/retry.js +5 -3
- package/dist/runner/retry.js.map +1 -1
- package/dist/runner/runner.d.ts +2 -14
- package/dist/runner/runner.d.ts.map +1 -1
- package/dist/runner/runner.js +11 -7
- package/dist/runner/runner.js.map +1 -1
- package/dist/runner/task-queue.d.ts +9 -11
- package/dist/runner/task-queue.d.ts.map +1 -1
- package/dist/runner/task-queue.js +79 -32
- package/dist/runner/task-queue.js.map +1 -1
- package/dist/storage/account-repository.d.ts +2 -15
- package/dist/storage/account-repository.d.ts.map +1 -1
- package/dist/storage/account-repository.js +65 -35
- package/dist/storage/account-repository.js.map +1 -1
- package/dist/storage/database.d.ts +3 -1
- package/dist/storage/database.d.ts.map +1 -1
- package/dist/storage/database.js +10 -4
- package/dist/storage/database.js.map +1 -1
- package/dist/storage/index.d.ts +3 -1
- package/dist/storage/index.d.ts.map +1 -1
- package/dist/storage/index.js +2 -0
- package/dist/storage/index.js.map +1 -1
- package/dist/storage/progress-repository.d.ts +12 -0
- package/dist/storage/progress-repository.d.ts.map +1 -0
- package/dist/storage/progress-repository.js +76 -0
- package/dist/storage/progress-repository.js.map +1 -0
- package/dist/storage/run-repository.d.ts +4 -10
- package/dist/storage/run-repository.d.ts.map +1 -1
- package/dist/storage/run-repository.js +4 -2
- package/dist/storage/run-repository.js.map +1 -1
- package/dist/storage/schema.d.ts.map +1 -1
- package/dist/storage/schema.js +19 -0
- package/dist/storage/schema.js.map +1 -1
- package/dist/storage/types.d.ts +15 -0
- package/dist/storage/types.d.ts.map +1 -0
- package/dist/storage/types.js +2 -0
- package/dist/storage/types.js.map +1 -0
- package/dist/transport/body.d.ts +15 -0
- package/dist/transport/body.d.ts.map +1 -0
- package/dist/transport/body.js +61 -0
- package/dist/transport/body.js.map +1 -0
- package/dist/transport/dispatcher-pool.d.ts +13 -0
- package/dist/transport/dispatcher-pool.d.ts.map +1 -0
- package/dist/transport/dispatcher-pool.js +47 -0
- package/dist/transport/dispatcher-pool.js.map +1 -0
- package/dist/transport/graphql.d.ts +1 -1
- package/dist/transport/graphql.d.ts.map +1 -1
- package/dist/transport/graphql.js +36 -19
- package/dist/transport/graphql.js.map +1 -1
- package/dist/transport/session.d.ts +8 -11
- package/dist/transport/session.d.ts.map +1 -1
- package/dist/transport/session.js +71 -26
- package/dist/transport/session.js.map +1 -1
- package/dist/transport/transaction-id.d.ts +2 -7
- package/dist/transport/transaction-id.d.ts.map +1 -1
- package/dist/transport/transaction-id.js +20 -16
- package/dist/transport/transaction-id.js.map +1 -1
- package/dist/transport/types.d.ts +28 -0
- package/dist/transport/types.d.ts.map +1 -0
- package/dist/transport/types.js +2 -0
- package/dist/transport/types.js.map +1 -0
- package/dist/utils/abort.d.ts +6 -0
- package/dist/utils/abort.d.ts.map +1 -0
- package/dist/utils/abort.js +45 -0
- package/dist/utils/abort.js.map +1 -0
- package/dist/utils/redact.d.ts +5 -0
- package/dist/utils/redact.d.ts.map +1 -0
- package/dist/utils/redact.js +42 -0
- package/dist/utils/redact.js.map +1 -0
- package/docs/architecture/overview.md +8 -4
- package/docs/product/specification.md +18 -8
- package/docs/security/data-boundary.md +3 -0
- package/package.json +5 -4
|
@@ -0,0 +1,47 @@
|
|
|
1
|
+
import { ManifestError } from "../domain/errors.js";
|
|
2
|
+
export const DEFAULT_CREDENTIAL_HOSTS = [
|
|
3
|
+
"x.com",
|
|
4
|
+
"www.x.com",
|
|
5
|
+
"api.x.com",
|
|
6
|
+
"twitter.com",
|
|
7
|
+
"www.twitter.com",
|
|
8
|
+
"api.twitter.com",
|
|
9
|
+
];
|
|
10
|
+
const GRAPHQL_PATH = /^\/(?:i\/api\/)?graphql\/[A-Za-z0-9_-]+\/[A-Za-z0-9_]+$/u;
|
|
11
|
+
export function allowedCredentialHosts(extraOrigins = []) {
|
|
12
|
+
const hosts = new Set(DEFAULT_CREDENTIAL_HOSTS);
|
|
13
|
+
for (const origin of extraOrigins) {
|
|
14
|
+
const host = originHost(origin);
|
|
15
|
+
if (host)
|
|
16
|
+
hosts.add(host);
|
|
17
|
+
}
|
|
18
|
+
return hosts;
|
|
19
|
+
}
|
|
20
|
+
export function assertCredentialDestination(value, extraOrigins = []) {
|
|
21
|
+
let parsed;
|
|
22
|
+
try {
|
|
23
|
+
parsed = new URL(value);
|
|
24
|
+
}
|
|
25
|
+
catch {
|
|
26
|
+
throw new ManifestError("Manifest endpoint is not a valid URL.");
|
|
27
|
+
}
|
|
28
|
+
if (parsed.protocol !== "https:")
|
|
29
|
+
throw new ManifestError(`Refusing non-HTTPS credential destination: ${parsed.origin}`);
|
|
30
|
+
if (!allowedCredentialHosts(extraOrigins).has(parsed.hostname))
|
|
31
|
+
throw new ManifestError(`Refusing untrusted credential destination: ${parsed.origin}`);
|
|
32
|
+
if (!GRAPHQL_PATH.test(parsed.pathname))
|
|
33
|
+
throw new ManifestError(`Refusing untrusted credential path: ${parsed.pathname}`);
|
|
34
|
+
return parsed.toString();
|
|
35
|
+
}
|
|
36
|
+
function originHost(value) {
|
|
37
|
+
const trimmed = value.trim();
|
|
38
|
+
if (!trimmed)
|
|
39
|
+
return undefined;
|
|
40
|
+
try {
|
|
41
|
+
return new URL(trimmed.includes("://") ? trimmed : `https://${trimmed}`).hostname;
|
|
42
|
+
}
|
|
43
|
+
catch {
|
|
44
|
+
return undefined;
|
|
45
|
+
}
|
|
46
|
+
}
|
|
47
|
+
//# sourceMappingURL=destinations.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"destinations.js","sourceRoot":"","sources":["../../src/manifest/destinations.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AAEpD,MAAM,CAAC,MAAM,wBAAwB,GAAG;IACtC,OAAO;IACP,WAAW;IACX,WAAW;IACX,aAAa;IACb,iBAAiB;IACjB,iBAAiB;CACT,CAAC;AAEX,MAAM,YAAY,GAAG,0DAA0D,CAAC;AAEhF,MAAM,UAAU,sBAAsB,CAAC,eAAkC,EAAE;IACzE,MAAM,KAAK,GAAG,IAAI,GAAG,CAAS,wBAAwB,CAAC,CAAC;IACxD,KAAK,MAAM,MAAM,IAAI,YAAY,EAAE,CAAC;QAClC,MAAM,IAAI,GAAG,UAAU,CAAC,MAAM,CAAC,CAAC;QAChC,IAAI,IAAI;YAAE,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;IAC5B,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC;AAED,MAAM,UAAU,2BAA2B,CAAC,KAAa,EAAE,eAAkC,EAAE;IAC7F,IAAI,MAAW,CAAC;IAChB,IAAI,CAAC;QACH,MAAM,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC;IAC1B,CAAC;IAAC,MAAM,CAAC;QACP,MAAM,IAAI,aAAa,CAAC,uCAAuC,CAAC,CAAC;IACnE,CAAC;IACD,IAAI,MAAM,CAAC,QAAQ,KAAK,QAAQ;QAC9B,MAAM,IAAI,aAAa,CAAC,8CAA8C,MAAM,CAAC,MAAM,EAAE,CAAC,CAAC;IACzF,IAAI,CAAC,sBAAsB,CAAC,YAAY,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,QAAQ,CAAC;QAC5D,MAAM,IAAI,aAAa,CAAC,8CAA8C,MAAM,CAAC,MAAM,EAAE,CAAC,CAAC;IACzF,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC;QACrC,MAAM,IAAI,aAAa,CAAC,uCAAuC,MAAM,CAAC,QAAQ,EAAE,CAAC,CAAC;IACpF,OAAO,MAAM,CAAC,QAAQ,EAAE,CAAC;AAC3B,CAAC;AAED,SAAS,UAAU,CAAC,KAAa;IAC/B,MAAM,OAAO,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC;IAC7B,IAAI,CAAC,OAAO;QAAE,OAAO,SAAS,CAAC;IAC/B,IAAI,CAAC;QACH,OAAO,IAAI,GAAG,CAAC,OAAO,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,WAAW,OAAO,EAAE,CAAC,CAAC,QAAQ,CAAC;IACpF,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,SAAS,CAAC;IACnB,CAAC;AACH,CAAC"}
|
package/dist/manifest/model.d.ts
CHANGED
|
@@ -1,3 +1,4 @@
|
|
|
1
|
-
import type { Manifest, ManifestPayload } from "../domain/manifest.js";
|
|
2
|
-
export
|
|
1
|
+
import type { Manifest, ManifestCreateOptions, ManifestPayload } from "../domain/manifest.js";
|
|
2
|
+
export type { ManifestCreateOptions } from "../domain/manifest.js";
|
|
3
|
+
export declare function createManifest(payload: ManifestPayload, options?: ManifestCreateOptions): Manifest;
|
|
3
4
|
//# sourceMappingURL=model.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"model.d.ts","sourceRoot":"","sources":["../../src/manifest/model.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,
|
|
1
|
+
{"version":3,"file":"model.d.ts","sourceRoot":"","sources":["../../src/manifest/model.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAGV,QAAQ,EACR,qBAAqB,EACrB,eAAe,EAChB,MAAM,uBAAuB,CAAC;AAK/B,YAAY,EAAE,qBAAqB,EAAE,MAAM,uBAAuB,CAAC;AAEnE,wBAAgB,cAAc,CAAC,OAAO,EAAE,eAAe,EAAE,OAAO,GAAE,qBAA0B,GAAG,QAAQ,CAiCtG"}
|
package/dist/manifest/model.js
CHANGED
|
@@ -1,6 +1,7 @@
|
|
|
1
1
|
import { ManifestError } from "../domain/errors.js";
|
|
2
2
|
import { asInteger, asString, isRecord } from "../utils/guards.js";
|
|
3
|
-
|
|
3
|
+
import { assertCredentialDestination } from "./destinations.js";
|
|
4
|
+
export function createManifest(payload, options = {}) {
|
|
4
5
|
if (!payload || !isRecord(payload.queryIds) || !isRecord(payload.endpoints)) {
|
|
5
6
|
throw new ManifestError("Manifest requires queryIds and endpoints objects.");
|
|
6
7
|
}
|
|
@@ -8,6 +9,10 @@ export function createManifest(payload) {
|
|
|
8
9
|
const endpoints = mapStrings(payload.endpoints);
|
|
9
10
|
if (!queryIds.search_timeline || !endpoints.search_timeline)
|
|
10
11
|
throw new ManifestError("Manifest requires a search endpoint.");
|
|
12
|
+
for (const [operation, endpoint] of Object.entries(endpoints)) {
|
|
13
|
+
const queryId = queryIds[operation] ?? "validation";
|
|
14
|
+
assertCredentialDestination(endpoint.includes("{query_id}") ? endpoint.replace("{query_id}", queryId) : endpoint, options.allowedOrigins ?? []);
|
|
15
|
+
}
|
|
11
16
|
const features = mapBooleans(payload.features);
|
|
12
17
|
const operationFeatures = mapBooleanMap(payload.operationFeatures);
|
|
13
18
|
const operationFieldToggles = mapBooleanMap(payload.operationFieldToggles);
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"model.js","sourceRoot":"","sources":["../../src/manifest/model.ts"],"names":[],"mappings":"
|
|
1
|
+
{"version":3,"file":"model.js","sourceRoot":"","sources":["../../src/manifest/model.ts"],"names":[],"mappings":"AAOA,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,SAAS,EAAE,QAAQ,EAAE,QAAQ,EAAE,MAAM,oBAAoB,CAAC;AACnE,OAAO,EAAE,2BAA2B,EAAE,MAAM,mBAAmB,CAAC;AAIhE,MAAM,UAAU,cAAc,CAAC,OAAwB,EAAE,UAAiC,EAAE;IAC1F,IAAI,CAAC,OAAO,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,SAAS,CAAC,EAAE,CAAC;QAC5E,MAAM,IAAI,aAAa,CAAC,mDAAmD,CAAC,CAAC;IAC/E,CAAC;IACD,MAAM,QAAQ,GAAG,UAAU,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC;IAC9C,MAAM,SAAS,GAAG,UAAU,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC;IAChD,IAAI,CAAC,QAAQ,CAAC,eAAe,IAAI,CAAC,SAAS,CAAC,eAAe;QACzD,MAAM,IAAI,aAAa,CAAC,sCAAsC,CAAC,CAAC;IAClE,KAAK,MAAM,CAAC,SAAS,EAAE,QAAQ,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,SAAS,CAAC,EAAE,CAAC;QAC9D,MAAM,OAAO,GAAG,QAAQ,CAAC,SAAS,CAAC,IAAI,YAAY,CAAC;QACpD,2BAA2B,CACzB,QAAQ,CAAC,QAAQ,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,YAAY,EAAE,OAAO,CAAC,CAAC,CAAC,CAAC,QAAQ,EACpF,OAAO,CAAC,cAAc,IAAI,EAAE,CAC7B,CAAC;IACJ,CAAC;IACD,MAAM,QAAQ,GAAG,WAAW,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC;IAC/C,MAAM,iBAAiB,GAAG,aAAa,CAAC,OAAO,CAAC,iBAAiB,CAAC,CAAC;IACnE,MAAM,qBAAqB,GAAG,aAAa,CAAC,OAAO,CAAC,qBAAqB,CAAC,CAAC;IAC3E,OAAO;QACL,OAAO,EAAE,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,IAAI,SAAS;QAC/C,QAAQ;QACR,SAAS;QACT,QAAQ;QACR,iBAAiB;QACjB,qBAAqB;QACrB,cAAc,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,SAAS,CAAC,OAAO,CAAC,cAAc,EAAE,EAAE,CAAC,CAAC;QAClE,WAAW,CAAC,SAAiB;YAC3B,OAAO,EAAE,GAAG,QAAQ,EAAE,GAAG,CAAC,iBAAiB,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC;QAClE,CAAC;QACD,eAAe,CAAC,SAAiB;YAC/B,OAAO,qBAAqB,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC;QAChD,CAAC;KACF,CAAC;AACJ,CAAC;AAED,SAAS,UAAU,CAAC,KAAc;IAChC,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAChC,OAAO,MAAM,CAAC,WAAW,CACvB,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,EAAE,KAAK,CAAC,EAAE,EAAE;QAC7C,MAAM,IAAI,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC;QAC7B,OAAO,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,EAAE,IAAI,CAAU,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IAC5C,CAAC,CAAC,CACH,CAAC;AACJ,CAAC;AAED,SAAS,WAAW,CAAC,KAAc;IACjC,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAChC,OAAO,MAAM,CAAC,WAAW,CACvB,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAA8B,EAAE,CAAC,OAAO,KAAK,CAAC,CAAC,CAAC,KAAK,SAAS,CAAC,CACnG,CAAC;AACJ,CAAC;AAED,SAAS,aAAa,CAAC,KAAc;IACnC,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC;QAAE,OAAO,EAAE,CAAC;IAChC,OAAO,MAAM,CAAC,WAAW,CAAC,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,GAAG,EAAE,KAAK,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,EAAE,WAAW,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;AACpG,CAAC"}
|
|
@@ -1,19 +1,26 @@
|
|
|
1
1
|
import type { ClientConfig } from "../config/types.js";
|
|
2
|
-
import type { Manifest, ManifestPayload } from "../domain/manifest.js";
|
|
2
|
+
import type { Manifest, ManifestPayload, ManifestScrapeOptions } from "../domain/manifest.js";
|
|
3
3
|
import type { ManifestRepository } from "../storage/manifest-repository.js";
|
|
4
|
-
|
|
4
|
+
export declare const LIVE_MANIFEST_CACHE_KEY = "xtrawl:live-web";
|
|
5
5
|
export declare class ManifestProvider {
|
|
6
6
|
private readonly config;
|
|
7
7
|
private readonly repository;
|
|
8
8
|
private readonly remoteFetch;
|
|
9
9
|
private readonly liveAuthToken?;
|
|
10
10
|
private readonly liveScrape;
|
|
11
|
-
private
|
|
11
|
+
private readonly local;
|
|
12
|
+
private active?;
|
|
13
|
+
private inFlightRefresh?;
|
|
14
|
+
private refreshFailedUntil;
|
|
12
15
|
private remoteRefreshAttempted;
|
|
13
16
|
constructor(config: ClientConfig, repository: ManifestRepository, remoteFetch?: (url: string) => Promise<unknown>, liveAuthToken?: string | undefined, liveScrape?: (base: ManifestPayload, options?: ManifestScrapeOptions) => Promise<ManifestPayload>);
|
|
14
17
|
getManifest(): Promise<Manifest>;
|
|
15
18
|
refreshLive(authToken?: string): Promise<Manifest>;
|
|
19
|
+
private loadRemote;
|
|
16
20
|
private fetchRemote;
|
|
17
21
|
private live;
|
|
22
|
+
private runLive;
|
|
23
|
+
private readCached;
|
|
24
|
+
private manifestFrom;
|
|
18
25
|
}
|
|
19
26
|
//# sourceMappingURL=provider.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"provider.d.ts","sourceRoot":"","sources":["../../src/manifest/provider.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AACvD,OAAO,KAAK,EAAE,QAAQ,EAAE,eAAe,EAAE,MAAM,uBAAuB,CAAC;
|
|
1
|
+
{"version":3,"file":"provider.d.ts","sourceRoot":"","sources":["../../src/manifest/provider.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,oBAAoB,CAAC;AACvD,OAAO,KAAK,EAAE,QAAQ,EAAE,eAAe,EAAE,qBAAqB,EAAE,MAAM,uBAAuB,CAAC;AAE9F,OAAO,KAAK,EAAE,kBAAkB,EAAE,MAAM,mCAAmC,CAAC;AAK5E,eAAO,MAAM,uBAAuB,oBAAoB,CAAC;AAGzD,qBAAa,gBAAgB;IAQzB,OAAO,CAAC,QAAQ,CAAC,MAAM;IACvB,OAAO,CAAC,QAAQ,CAAC,UAAU;IAC3B,OAAO,CAAC,QAAQ,CAAC,WAAW;IAC5B,OAAO,CAAC,QAAQ,CAAC,aAAa,CAAC;IAC/B,OAAO,CAAC,QAAQ,CAAC,UAAU;IAX7B,OAAO,CAAC,QAAQ,CAAC,KAAK,CAAW;IACjC,OAAO,CAAC,MAAM,CAAC,CAAW;IAC1B,OAAO,CAAC,eAAe,CAAC,CAAoB;IAC5C,OAAO,CAAC,kBAAkB,CAAK;IAC/B,OAAO,CAAC,sBAAsB,CAAS;gBAGpB,MAAM,EAAE,YAAY,EACpB,UAAU,EAAE,kBAAkB,EAC9B,WAAW,GAAE,CAAC,GAAG,EAAE,MAAM,KAAK,OAAO,CAAC,OAAO,CAAoB,EACjE,aAAa,CAAC,EAAE,MAAM,YAAA,EACtB,UAAU,GAAE,CAC3B,IAAI,EAAE,eAAe,EACrB,OAAO,CAAC,EAAE,qBAAqB,KAC5B,OAAO,CAAC,eAAe,CAAyB;IAK1C,WAAW,IAAI,OAAO,CAAC,QAAQ,CAAC;IAmBhC,WAAW,CAAC,SAAS,CAAC,EAAE,MAAM,GAAG,OAAO,CAAC,QAAQ,CAAC;YAIjD,UAAU;YAqBV,WAAW;YAQX,IAAI;YAaJ,OAAO;IAuBrB,OAAO,CAAC,UAAU;IAUlB,OAAO,CAAC,YAAY;CAGrB"}
|
|
@@ -2,13 +2,18 @@ import { ManifestError } from "../domain/errors.js";
|
|
|
2
2
|
import { DEFAULT_MANIFEST } from "./default-manifest.js";
|
|
3
3
|
import { createManifest } from "./model.js";
|
|
4
4
|
import { scrapeManifestFromWeb } from "./scraper.js";
|
|
5
|
+
export const LIVE_MANIFEST_CACHE_KEY = "xtrawl:live-web";
|
|
6
|
+
const REFRESH_BACKOFF_MS = 30_000;
|
|
5
7
|
export class ManifestProvider {
|
|
6
8
|
config;
|
|
7
9
|
repository;
|
|
8
10
|
remoteFetch;
|
|
9
11
|
liveAuthToken;
|
|
10
12
|
liveScrape;
|
|
11
|
-
|
|
13
|
+
local;
|
|
14
|
+
active;
|
|
15
|
+
inFlightRefresh;
|
|
16
|
+
refreshFailedUntil = 0;
|
|
12
17
|
remoteRefreshAttempted = false;
|
|
13
18
|
constructor(config, repository, remoteFetch = defaultFetchJson, liveAuthToken, liveScrape = scrapeManifestFromWeb) {
|
|
14
19
|
this.config = config;
|
|
@@ -16,66 +21,113 @@ export class ManifestProvider {
|
|
|
16
21
|
this.remoteFetch = remoteFetch;
|
|
17
22
|
this.liveAuthToken = liveAuthToken;
|
|
18
23
|
this.liveScrape = liveScrape;
|
|
24
|
+
this.local = this.manifestFrom(DEFAULT_MANIFEST);
|
|
19
25
|
}
|
|
20
26
|
async getManifest() {
|
|
21
|
-
|
|
22
|
-
|
|
23
|
-
|
|
24
|
-
|
|
25
|
-
return this.live(local, false);
|
|
26
|
-
}
|
|
27
|
+
if (this.active)
|
|
28
|
+
return this.active;
|
|
29
|
+
if (this.config.manifestScrapeOnInit)
|
|
30
|
+
return this.live(false);
|
|
27
31
|
if (this.config.manifestUrl) {
|
|
28
|
-
|
|
29
|
-
|
|
30
|
-
|
|
31
|
-
|
|
32
|
-
}
|
|
33
|
-
catch {
|
|
34
|
-
const stale = this.repository.get(this.config.manifestUrl, true);
|
|
35
|
-
if (stale)
|
|
36
|
-
return createManifest(stale);
|
|
37
|
-
}
|
|
32
|
+
const remote = await this.loadRemote(this.config.manifestUrl);
|
|
33
|
+
if (remote) {
|
|
34
|
+
this.active = remote;
|
|
35
|
+
return remote;
|
|
38
36
|
}
|
|
39
|
-
|
|
40
|
-
|
|
41
|
-
|
|
37
|
+
}
|
|
38
|
+
const cachedLive = this.readCached(LIVE_MANIFEST_CACHE_KEY);
|
|
39
|
+
if (cachedLive) {
|
|
40
|
+
this.active = cachedLive;
|
|
41
|
+
return cachedLive;
|
|
42
|
+
}
|
|
43
|
+
this.active = this.local;
|
|
44
|
+
return this.local;
|
|
45
|
+
}
|
|
46
|
+
async refreshLive(authToken) {
|
|
47
|
+
return this.live(true, authToken);
|
|
48
|
+
}
|
|
49
|
+
async loadRemote(url) {
|
|
50
|
+
if (this.config.manifestUpdateOnInit && !this.remoteRefreshAttempted) {
|
|
51
|
+
this.remoteRefreshAttempted = true;
|
|
42
52
|
try {
|
|
43
|
-
return await this.fetchRemote(
|
|
53
|
+
return await this.fetchRemote(url);
|
|
44
54
|
}
|
|
45
|
-
catch {
|
|
46
|
-
|
|
55
|
+
catch (error) {
|
|
56
|
+
if (error instanceof ManifestError)
|
|
57
|
+
throw error;
|
|
58
|
+
const stale = this.readCached(url, true);
|
|
47
59
|
if (stale)
|
|
48
|
-
return
|
|
60
|
+
return stale;
|
|
49
61
|
}
|
|
50
62
|
}
|
|
51
|
-
|
|
52
|
-
|
|
53
|
-
|
|
54
|
-
|
|
63
|
+
const cached = this.readCached(url);
|
|
64
|
+
if (cached)
|
|
65
|
+
return cached;
|
|
66
|
+
try {
|
|
67
|
+
return await this.fetchRemote(url);
|
|
68
|
+
}
|
|
69
|
+
catch (error) {
|
|
70
|
+
if (error instanceof ManifestError)
|
|
71
|
+
throw error;
|
|
72
|
+
return this.readCached(url, true);
|
|
73
|
+
}
|
|
55
74
|
}
|
|
56
75
|
async fetchRemote(url) {
|
|
57
76
|
const remote = await this.remoteFetch(url);
|
|
58
77
|
const payload = normalizePayload(remote);
|
|
78
|
+
const manifest = this.manifestFrom(payload);
|
|
59
79
|
this.repository.set(url, payload, this.config.manifestTtlMs);
|
|
60
|
-
return
|
|
80
|
+
return manifest;
|
|
81
|
+
}
|
|
82
|
+
async live(strict, authToken) {
|
|
83
|
+
if (this.inFlightRefresh)
|
|
84
|
+
return this.inFlightRefresh;
|
|
85
|
+
if (Date.now() < this.refreshFailedUntil) {
|
|
86
|
+
if (strict)
|
|
87
|
+
throw new ManifestError("Live manifest refresh is in backoff.");
|
|
88
|
+
return this.active ?? this.local;
|
|
89
|
+
}
|
|
90
|
+
const pending = this.runLive(strict, authToken).finally(() => {
|
|
91
|
+
if (this.inFlightRefresh === pending)
|
|
92
|
+
this.inFlightRefresh = undefined;
|
|
93
|
+
});
|
|
94
|
+
this.inFlightRefresh = pending;
|
|
95
|
+
return pending;
|
|
61
96
|
}
|
|
62
|
-
async
|
|
97
|
+
async runLive(strict, authToken) {
|
|
63
98
|
try {
|
|
64
99
|
const token = authToken ?? this.liveAuthToken;
|
|
65
100
|
const payload = await this.liveScrape(DEFAULT_MANIFEST, {
|
|
66
101
|
...(token ? { authToken: token } : {}),
|
|
67
102
|
});
|
|
103
|
+
this.active = this.manifestFrom(payload);
|
|
104
|
+
this.repository.set(this.config.manifestUrl ?? LIVE_MANIFEST_CACHE_KEY, payload, this.config.manifestTtlMs);
|
|
68
105
|
if (this.config.manifestUrl)
|
|
69
|
-
this.repository.set(
|
|
70
|
-
this.
|
|
71
|
-
return this.
|
|
106
|
+
this.repository.set(LIVE_MANIFEST_CACHE_KEY, payload, this.config.manifestTtlMs);
|
|
107
|
+
this.refreshFailedUntil = 0;
|
|
108
|
+
return this.active;
|
|
72
109
|
}
|
|
73
110
|
catch (error) {
|
|
111
|
+
this.refreshFailedUntil = Date.now() + REFRESH_BACKOFF_MS;
|
|
74
112
|
if (strict)
|
|
75
113
|
throw new ManifestError(`Live manifest refresh failed: ${String(error)}`);
|
|
76
|
-
return local;
|
|
114
|
+
return this.active ?? this.local;
|
|
77
115
|
}
|
|
78
116
|
}
|
|
117
|
+
readCached(key, allowExpired = false) {
|
|
118
|
+
const payload = this.repository.get(key, allowExpired);
|
|
119
|
+
if (!payload)
|
|
120
|
+
return undefined;
|
|
121
|
+
try {
|
|
122
|
+
return this.manifestFrom(payload);
|
|
123
|
+
}
|
|
124
|
+
catch {
|
|
125
|
+
return undefined;
|
|
126
|
+
}
|
|
127
|
+
}
|
|
128
|
+
manifestFrom(payload) {
|
|
129
|
+
return createManifest(payload, { allowedOrigins: this.config.allowedManifestOrigins });
|
|
130
|
+
}
|
|
79
131
|
}
|
|
80
132
|
async function defaultFetchJson(url) {
|
|
81
133
|
const response = await fetch(url);
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"provider.js","sourceRoot":"","sources":["../../src/manifest/provider.ts"],"names":[],"mappings":"AAEA,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AAEpD,OAAO,EAAE,gBAAgB,EAAE,MAAM,uBAAuB,CAAC;AACzD,OAAO,EAAE,cAAc,EAAE,MAAM,YAAY,CAAC;AAC5C,OAAO,EAAE,qBAAqB,EAAE,MAAM,cAAc,CAAC;
|
|
1
|
+
{"version":3,"file":"provider.js","sourceRoot":"","sources":["../../src/manifest/provider.ts"],"names":[],"mappings":"AAEA,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AAEpD,OAAO,EAAE,gBAAgB,EAAE,MAAM,uBAAuB,CAAC;AACzD,OAAO,EAAE,cAAc,EAAE,MAAM,YAAY,CAAC;AAC5C,OAAO,EAAE,qBAAqB,EAAE,MAAM,cAAc,CAAC;AAErD,MAAM,CAAC,MAAM,uBAAuB,GAAG,iBAAiB,CAAC;AACzD,MAAM,kBAAkB,GAAG,MAAM,CAAC;AAElC,MAAM,OAAO,gBAAgB;IAQR;IACA;IACA;IACA;IACA;IAXF,KAAK,CAAW;IACzB,MAAM,CAAY;IAClB,eAAe,CAAqB;IACpC,kBAAkB,GAAG,CAAC,CAAC;IACvB,sBAAsB,GAAG,KAAK,CAAC;IAEvC,YACmB,MAAoB,EACpB,UAA8B,EAC9B,cAAiD,gBAAgB,EACjE,aAAsB,EACtB,aAGe,qBAAqB;QAPpC,WAAM,GAAN,MAAM,CAAc;QACpB,eAAU,GAAV,UAAU,CAAoB;QAC9B,gBAAW,GAAX,WAAW,CAAsD;QACjE,kBAAa,GAAb,aAAa,CAAS;QACtB,eAAU,GAAV,UAAU,CAG0B;QAErD,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,YAAY,CAAC,gBAAgB,CAAC,CAAC;IACnD,CAAC;IAEM,KAAK,CAAC,WAAW;QACtB,IAAI,IAAI,CAAC,MAAM;YAAE,OAAO,IAAI,CAAC,MAAM,CAAC;QACpC,IAAI,IAAI,CAAC,MAAM,CAAC,oBAAoB;YAAE,OAAO,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAC9D,IAAI,IAAI,CAAC,MAAM,CAAC,WAAW,EAAE,CAAC;YAC5B,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,WAAW,CAAC,CAAC;YAC9D,IAAI,MAAM,EAAE,CAAC;gBACX,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;gBACrB,OAAO,MAAM,CAAC;YAChB,CAAC;QACH,CAAC;QACD,MAAM,UAAU,GAAG,IAAI,CAAC,UAAU,CAAC,uBAAuB,CAAC,CAAC;QAC5D,IAAI,UAAU,EAAE,CAAC;YACf,IAAI,CAAC,MAAM,GAAG,UAAU,CAAC;YACzB,OAAO,UAAU,CAAC;QACpB,CAAC;QACD,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC;QACzB,OAAO,IAAI,CAAC,KAAK,CAAC;IACpB,CAAC;IAEM,KAAK,CAAC,WAAW,CAAC,SAAkB;QACzC,OAAO,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,SAAS,CAAC,CAAC;IACpC,CAAC;IAEO,KAAK,CAAC,UAAU,CAAC,GAAW;QAClC,IAAI,IAAI,CAAC,MAAM,CAAC,oBAAoB,IAAI,CAAC,IAAI,CAAC,sBAAsB,EAAE,CAAC;YACrE,IAAI,CAAC,sBAAsB,GAAG,IAAI,CAAC;YACnC,IAAI,CAAC;gBACH,OAAO,MAAM,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;YACrC,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,IAAI,KAAK,YAAY,aAAa;oBAAE,MAAM,KAAK,CAAC;gBAChD,MAAM,KAAK,GAAG,IAAI,CAAC,UAAU,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;gBACzC,IAAI,KAAK;oBAAE,OAAO,KAAK,CAAC;YAC1B,CAAC;QACH,CAAC;QACD,MAAM,MAAM,GAAG,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC;QACpC,IAAI,MAAM;YAAE,OAAO,MAAM,CAAC;QAC1B,IAAI,CAAC;YACH,OAAO,MAAM,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;QACrC,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,IAAI,KAAK,YAAY,aAAa;gBAAE,MAAM,KAAK,CAAC;YAChD,OAAO,IAAI,CAAC,UAAU,CAAC,GAAG,EAAE,IAAI,CAAC,CAAC;QACpC,CAAC;IACH,CAAC;IAEO,KAAK,CAAC,WAAW,CAAC,GAAW;QACnC,MAAM,MAAM,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;QAC3C,MAAM,OAAO,GAAG,gBAAgB,CAAC,MAAM,CAAC,CAAC;QACzC,MAAM,QAAQ,GAAG,IAAI,CAAC,YAAY,CAAC,OAAO,CAAC,CAAC;QAC5C,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,GAAG,EAAE,OAAO,EAAE,IAAI,CAAC,MAAM,CAAC,aAAa,CAAC,CAAC;QAC7D,OAAO,QAAQ,CAAC;IAClB,CAAC;IAEO,KAAK,CAAC,IAAI,CAAC,MAAe,EAAE,SAAkB;QACpD,IAAI,IAAI,CAAC,eAAe;YAAE,OAAO,IAAI,CAAC,eAAe,CAAC;QACtD,IAAI,IAAI,CAAC,GAAG,EAAE,GAAG,IAAI,CAAC,kBAAkB,EAAE,CAAC;YACzC,IAAI,MAAM;gBAAE,MAAM,IAAI,aAAa,CAAC,sCAAsC,CAAC,CAAC;YAC5E,OAAO,IAAI,CAAC,MAAM,IAAI,IAAI,CAAC,KAAK,CAAC;QACnC,CAAC;QACD,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC,OAAO,CAAC,GAAG,EAAE;YAC3D,IAAI,IAAI,CAAC,eAAe,KAAK,OAAO;gBAAE,IAAI,CAAC,eAAe,GAAG,SAAS,CAAC;QACzE,CAAC,CAAC,CAAC;QACH,IAAI,CAAC,eAAe,GAAG,OAAO,CAAC;QAC/B,OAAO,OAAO,CAAC;IACjB,CAAC;IAEO,KAAK,CAAC,OAAO,CAAC,MAAe,EAAE,SAAkB;QACvD,IAAI,CAAC;YACH,MAAM,KAAK,GAAG,SAAS,IAAI,IAAI,CAAC,aAAa,CAAC;YAC9C,MAAM,OAAO,GAAG,MAAM,IAAI,CAAC,UAAU,CAAC,gBAAgB,EAAE;gBACtD,GAAG,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,SAAS,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;aACvC,CAAC,CAAC;YACH,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC,YAAY,CAAC,OAAO,CAAC,CAAC;YACzC,IAAI,CAAC,UAAU,CAAC,GAAG,CACjB,IAAI,CAAC,MAAM,CAAC,WAAW,IAAI,uBAAuB,EAClD,OAAO,EACP,IAAI,CAAC,MAAM,CAAC,aAAa,CAC1B,CAAC;YACF,IAAI,IAAI,CAAC,MAAM,CAAC,WAAW;gBACzB,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,uBAAuB,EAAE,OAAO,EAAE,IAAI,CAAC,MAAM,CAAC,aAAa,CAAC,CAAC;YACnF,IAAI,CAAC,kBAAkB,GAAG,CAAC,CAAC;YAC5B,OAAO,IAAI,CAAC,MAAM,CAAC;QACrB,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,IAAI,CAAC,kBAAkB,GAAG,IAAI,CAAC,GAAG,EAAE,GAAG,kBAAkB,CAAC;YAC1D,IAAI,MAAM;gBAAE,MAAM,IAAI,aAAa,CAAC,iCAAiC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;YACtF,OAAO,IAAI,CAAC,MAAM,IAAI,IAAI,CAAC,KAAK,CAAC;QACnC,CAAC;IACH,CAAC;IAEO,UAAU,CAAC,GAAW,EAAE,YAAY,GAAG,KAAK;QAClD,MAAM,OAAO,GAAG,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,GAAG,EAAE,YAAY,CAAC,CAAC;QACvD,IAAI,CAAC,OAAO;YAAE,OAAO,SAAS,CAAC;QAC/B,IAAI,CAAC;YACH,OAAO,IAAI,CAAC,YAAY,CAAC,OAAO,CAAC,CAAC;QACpC,CAAC;QAAC,MAAM,CAAC;YACP,OAAO,SAAS,CAAC;QACnB,CAAC;IACH,CAAC;IAEO,YAAY,CAAC,OAAwB;QAC3C,OAAO,cAAc,CAAC,OAAO,EAAE,EAAE,cAAc,EAAE,IAAI,CAAC,MAAM,CAAC,sBAAsB,EAAE,CAAC,CAAC;IACzF,CAAC;CACF;AAED,KAAK,UAAU,gBAAgB,CAAC,GAAW;IACzC,MAAM,QAAQ,GAAG,MAAM,KAAK,CAAC,GAAG,CAAC,CAAC;IAClC,IAAI,CAAC,QAAQ,CAAC,EAAE;QAAE,MAAM,IAAI,KAAK,CAAC,qCAAqC,QAAQ,CAAC,MAAM,EAAE,CAAC,CAAC;IAC1F,OAAO,QAAQ,CAAC,IAAI,EAAsB,CAAC;AAC7C,CAAC;AAED,SAAS,gBAAgB,CAAC,KAAc;IACtC,IAAI,CAAC,KAAK,IAAI,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC;QAC7D,MAAM,IAAI,aAAa,CAAC,mCAAmC,CAAC,CAAC;IAC/D,MAAM,SAAS,GAAG,KAAiC,CAAC;IACpD,IAAI,CAAC,SAAS,CAAC,QAAQ,IAAI,CAAC,SAAS,CAAC,SAAS;QAC7C,MAAM,IAAI,aAAa,CAAC,mDAAmD,CAAC,CAAC;IAC/E,OAAO,SAA4B,CAAC;AACtC,CAAC"}
|
|
@@ -1,9 +1,5 @@
|
|
|
1
|
-
import type { ManifestPayload } from "../domain/manifest.js";
|
|
2
|
-
export
|
|
3
|
-
readonly authToken?: string;
|
|
4
|
-
readonly fetcher?: typeof fetch;
|
|
5
|
-
readonly maxScripts?: number;
|
|
6
|
-
}
|
|
1
|
+
import type { ManifestPayload, ManifestScrapeOptions } from "../domain/manifest.js";
|
|
2
|
+
export type { ManifestScrapeOptions } from "../domain/manifest.js";
|
|
7
3
|
export declare function extractManifestFromJavascript(source: string, fallback: ManifestPayload): ManifestPayload;
|
|
8
4
|
export declare function extractOperationQueryIdsFromJavascript(source: string): Record<string, string>;
|
|
9
5
|
export declare function extractOperationFeaturesFromJavascript(source: string): Record<string, Record<string, boolean>>;
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"scraper.d.ts","sourceRoot":"","sources":["../../src/manifest/scraper.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,eAAe,EAAE,
|
|
1
|
+
{"version":3,"file":"scraper.d.ts","sourceRoot":"","sources":["../../src/manifest/scraper.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,eAAe,EAAE,qBAAqB,EAAE,MAAM,uBAAuB,CAAC;AAmBpF,YAAY,EAAE,qBAAqB,EAAE,MAAM,uBAAuB,CAAC;AAEnE,wBAAgB,6BAA6B,CAAC,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,eAAe,GAAG,eAAe,CAWxG;AAED,wBAAgB,sCAAsC,CAAC,MAAM,EAAE,MAAM,GAAG,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAe7F;AAED,wBAAgB,sCAAsC,CACpD,MAAM,EAAE,MAAM,GACb,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC,CAmBzC;AAED,wBAAsB,qBAAqB,CACzC,IAAI,EAAE,eAAe,EACrB,OAAO,GAAE,qBAA0B,GAClC,OAAO,CAAC,eAAe,CAAC,CA+C1B"}
|
package/dist/manifest/scraper.js
CHANGED
|
@@ -1,3 +1,4 @@
|
|
|
1
|
+
import { AUXILIARY_REQUEST_TIMEOUT_MS, cancelBody, DEFAULT_MAX_RESPONSE_BYTES, readResponseText, } from "../transport/body.js";
|
|
1
2
|
import { asString } from "../utils/guards.js";
|
|
2
3
|
const OPERATION_NAMES = {
|
|
3
4
|
SearchTimeline: "search_timeline",
|
|
@@ -64,24 +65,34 @@ export async function scrapeManifestFromWeb(base, options = {}) {
|
|
|
64
65
|
...scriptHeaders,
|
|
65
66
|
...(options.authToken ? { Cookie: `auth_token=${options.authToken}` } : {}),
|
|
66
67
|
};
|
|
67
|
-
const
|
|
68
|
-
|
|
69
|
-
|
|
70
|
-
|
|
71
|
-
const
|
|
72
|
-
|
|
73
|
-
|
|
74
|
-
|
|
75
|
-
|
|
76
|
-
|
|
77
|
-
|
|
68
|
+
const html = await fetchBoundedText(fetcher, "https://x.com/home", {
|
|
69
|
+
headers: pageHeaders,
|
|
70
|
+
redirect: "error",
|
|
71
|
+
});
|
|
72
|
+
const seen = new Set();
|
|
73
|
+
const scripts = [];
|
|
74
|
+
for (const match of html.matchAll(/<script[^>]+src=["']([^"']+)["']/giu)) {
|
|
75
|
+
const raw = match[1];
|
|
76
|
+
if (!raw)
|
|
77
|
+
continue;
|
|
78
|
+
const url = safeScriptUrl(raw);
|
|
79
|
+
if (!url || seen.has(url))
|
|
80
|
+
continue;
|
|
81
|
+
seen.add(url);
|
|
82
|
+
scripts.push(url);
|
|
83
|
+
}
|
|
84
|
+
scripts.sort((left, right) => scriptPriority(left) - scriptPriority(right));
|
|
85
|
+
const selected = scripts.slice(0, options.maxScripts ?? 20);
|
|
78
86
|
const discovered = {};
|
|
79
87
|
const discoveredFeatures = {};
|
|
80
|
-
for (const url of
|
|
81
|
-
|
|
82
|
-
|
|
88
|
+
for (const url of selected) {
|
|
89
|
+
let source;
|
|
90
|
+
try {
|
|
91
|
+
source = await fetchBoundedText(fetcher, url, { headers: scriptHeaders, redirect: "follow" });
|
|
92
|
+
}
|
|
93
|
+
catch {
|
|
83
94
|
continue;
|
|
84
|
-
|
|
95
|
+
}
|
|
85
96
|
Object.assign(discovered, extractOperationQueryIdsFromJavascript(source));
|
|
86
97
|
Object.assign(discoveredFeatures, extractOperationFeaturesFromJavascript(source));
|
|
87
98
|
if (Object.keys(discovered).length === Object.keys(OPERATION_NAMES).length)
|
|
@@ -97,6 +108,26 @@ export async function scrapeManifestFromWeb(base, options = {}) {
|
|
|
97
108
|
operationFeatures: mergeOperationFeatures(base.operationFeatures, discoveredFeatures),
|
|
98
109
|
};
|
|
99
110
|
}
|
|
111
|
+
async function fetchBoundedText(fetcher, url, init) {
|
|
112
|
+
const timeout = new AbortController();
|
|
113
|
+
const timer = setTimeout(() => timeout.abort(), AUXILIARY_REQUEST_TIMEOUT_MS);
|
|
114
|
+
const signal = init.signal ? AbortSignal.any([init.signal, timeout.signal]) : timeout.signal;
|
|
115
|
+
try {
|
|
116
|
+
const response = await fetcher(url, { ...init, signal, redirect: init.redirect ?? "follow" });
|
|
117
|
+
try {
|
|
118
|
+
if (!response.ok)
|
|
119
|
+
throw new Error(`Manifest fetch failed with status ${response.status}`);
|
|
120
|
+
return await readResponseText(response, { signal, maxBytes: DEFAULT_MAX_RESPONSE_BYTES });
|
|
121
|
+
}
|
|
122
|
+
catch (error) {
|
|
123
|
+
await cancelBody(response);
|
|
124
|
+
throw error;
|
|
125
|
+
}
|
|
126
|
+
}
|
|
127
|
+
finally {
|
|
128
|
+
clearTimeout(timer);
|
|
129
|
+
}
|
|
130
|
+
}
|
|
100
131
|
function mergeOperationFeatures(base, discovered) {
|
|
101
132
|
const out = {};
|
|
102
133
|
if (base)
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"scraper.js","sourceRoot":"","sources":["../../src/manifest/scraper.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,QAAQ,EAAE,MAAM,oBAAoB,CAAC;AAE9C,MAAM,eAAe,GAA2B;IAC9C,cAAc,EAAE,iBAAiB;IACjC,gBAAgB,EAAE,yBAAyB;IAC3C,UAAU,EAAE,kBAAkB;IAC9B,SAAS,EAAE,WAAW;IACtB,SAAS,EAAE,WAAW;IACtB,qBAAqB,EAAE,oBAAoB;IAC3C,mBAAmB,EAAE,cAAc;CACpC,CAAC;
|
|
1
|
+
{"version":3,"file":"scraper.js","sourceRoot":"","sources":["../../src/manifest/scraper.ts"],"names":[],"mappings":"AACA,OAAO,EACL,4BAA4B,EAC5B,UAAU,EACV,0BAA0B,EAC1B,gBAAgB,GACjB,MAAM,sBAAsB,CAAC;AAC9B,OAAO,EAAE,QAAQ,EAAE,MAAM,oBAAoB,CAAC;AAE9C,MAAM,eAAe,GAA2B;IAC9C,cAAc,EAAE,iBAAiB;IACjC,gBAAgB,EAAE,yBAAyB;IAC3C,UAAU,EAAE,kBAAkB;IAC9B,SAAS,EAAE,WAAW;IACtB,SAAS,EAAE,WAAW;IACtB,qBAAqB,EAAE,oBAAoB;IAC3C,mBAAmB,EAAE,cAAc;CACpC,CAAC;AAIF,MAAM,UAAU,6BAA6B,CAAC,MAAc,EAAE,QAAyB;IACrF,MAAM,iBAAiB,GAAG,sCAAsC,CAAC,MAAM,CAAC,CAAC;IACzE,OAAO;QACL,GAAG,QAAQ;QACX,QAAQ,EAAE;YACR,GAAG,QAAQ,CAAC,QAAQ;YACpB,GAAG,sCAAsC,CAAC,MAAM,CAAC;SAClD;QACD,iBAAiB,EAAE,sBAAsB,CAAC,QAAQ,CAAC,iBAAiB,EAAE,iBAAiB,CAAC;QACxF,OAAO,EAAE,QAAQ,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,MAAM;KAC9C,CAAC;AACJ,CAAC;AAED,MAAM,UAAU,sCAAsC,CAAC,MAAc;IACnE,MAAM,QAAQ,GAA2B,EAAE,CAAC;IAC5C,KAAK,MAAM,CAAC,aAAa,EAAE,GAAG,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,eAAe,CAAC,EAAE,CAAC;QACnE,MAAM,UAAU,GAAG,IAAI,MAAM,CAC3B,sEAAsE,aAAa,MAAM,EACzF,GAAG,CACJ,CAAC;QACF,MAAM,cAAc,GAAG,IAAI,MAAM,CAC/B,6BAA6B,aAAa,+CAA+C,EACzF,GAAG,CACJ,CAAC;QACF,MAAM,EAAE,GAAG,MAAM,CAAC,KAAK,CAAC,UAAU,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,MAAM,CAAC,KAAK,CAAC,cAAc,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;QAC9E,IAAI,EAAE;YAAE,QAAQ,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC;IAC7B,CAAC;IACD,OAAO,QAAQ,CAAC;AAClB,CAAC;AAED,MAAM,UAAU,sCAAsC,CACpD,MAAc;IAEd,MAAM,GAAG,GAA4C,EAAE,CAAC;IACxD,KAAK,MAAM,CAAC,aAAa,EAAE,GAAG,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,eAAe,CAAC,EAAE,CAAC;QACnE,MAAM,MAAM,GAAG,IAAI,MAAM,CAAC,6BAA6B,aAAa,MAAM,EAAE,IAAI,CAAC,CAAC;QAClF,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QAClC,IAAI,CAAC,KAAK;YAAE,SAAS;QACrB,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC;QAC1B,MAAM,IAAI,GAAG,MAAM,CAAC,OAAO,CAAC,gBAAgB,EAAE,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC;QACvE,MAAM,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,KAAK,EAAE,IAAI,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,KAAK,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,KAAK,GAAG,MAAM,CAAC,CAAC;QACnG,MAAM,YAAY,GAAG,MAAM,CAAC,OAAO,CAAC,mBAAmB,CAAC,CAAC;QACzD,IAAI,YAAY,GAAG,CAAC;YAAE,SAAS;QAC/B,MAAM,UAAU,GAAG,MAAM,CAAC,OAAO,CAAC,GAAG,EAAE,YAAY,CAAC,CAAC;QACrD,IAAI,UAAU,GAAG,CAAC;YAAE,SAAS;QAC7B,MAAM,KAAK,GAAG,CAAC,GAAG,MAAM,CAAC,KAAK,CAAC,YAAY,EAAE,UAAU,CAAC,CAAC,QAAQ,CAAC,oBAAoB,CAAC,CAAC;aACrF,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;aACtB,MAAM,CAAC,CAAC,IAAI,EAAkB,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC;QACnD,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC;YAAE,GAAG,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,WAAW,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC;IAC1F,CAAC;IACD,OAAO,GAAG,CAAC;AACb,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,qBAAqB,CACzC,IAAqB,EACrB,UAAiC,EAAE;IAEnC,MAAM,OAAO,GAAG,OAAO,CAAC,OAAO,IAAI,KAAK,CAAC;IACzC,MAAM,aAAa,GAAG;QACpB,YAAY,EAAE,aAAa;KAC5B,CAAC;IACF,MAAM,WAAW,GAAG;QAClB,GAAG,aAAa;QAChB,GAAG,CAAC,OAAO,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,cAAc,OAAO,CAAC,SAAS,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KAC5E,CAAC;IACF,MAAM,IAAI,GAAG,MAAM,gBAAgB,CAAC,OAAO,EAAE,oBAAoB,EAAE;QACjE,OAAO,EAAE,WAAW;QACpB,QAAQ,EAAE,OAAO;KAClB,CAAC,CAAC;IACH,MAAM,IAAI,GAAG,IAAI,GAAG,EAAU,CAAC;IAC/B,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,KAAK,MAAM,KAAK,IAAI,IAAI,CAAC,QAAQ,CAAC,qCAAqC,CAAC,EAAE,CAAC;QACzE,MAAM,GAAG,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC;QACrB,IAAI,CAAC,GAAG;YAAE,SAAS;QACnB,MAAM,GAAG,GAAG,aAAa,CAAC,GAAG,CAAC,CAAC;QAC/B,IAAI,CAAC,GAAG,IAAI,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC;YAAE,SAAS;QACpC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QACd,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IACpB,CAAC;IACD,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,cAAc,CAAC,IAAI,CAAC,GAAG,cAAc,CAAC,KAAK,CAAC,CAAC,CAAC;IAC5E,MAAM,QAAQ,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,UAAU,IAAI,EAAE,CAAC,CAAC;IAC5D,MAAM,UAAU,GAA2B,EAAE,CAAC;IAC9C,MAAM,kBAAkB,GAA4C,EAAE,CAAC;IACvE,KAAK,MAAM,GAAG,IAAI,QAAQ,EAAE,CAAC;QAC3B,IAAI,MAAc,CAAC;QACnB,IAAI,CAAC;YACH,MAAM,GAAG,MAAM,gBAAgB,CAAC,OAAO,EAAE,GAAG,EAAE,EAAE,OAAO,EAAE,aAAa,EAAE,QAAQ,EAAE,QAAQ,EAAE,CAAC,CAAC;QAChG,CAAC;QAAC,MAAM,CAAC;YACP,SAAS;QACX,CAAC;QACD,MAAM,CAAC,MAAM,CAAC,UAAU,EAAE,sCAAsC,CAAC,MAAM,CAAC,CAAC,CAAC;QAC1E,MAAM,CAAC,MAAM,CAAC,kBAAkB,EAAE,sCAAsC,CAAC,MAAM,CAAC,CAAC,CAAC;QAClF,IAAI,MAAM,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,MAAM,KAAK,MAAM,CAAC,IAAI,CAAC,eAAe,CAAC,CAAC,MAAM;YAAE,MAAM;IACpF,CAAC;IACD,IAAI,MAAM,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACzC,MAAM,IAAI,KAAK,CAAC,sEAAsE,CAAC,CAAC;IAC1F,CAAC;IACD,OAAO;QACL,GAAG,IAAI;QACP,OAAO,EAAE,UAAU;QACnB,QAAQ,EAAE,EAAE,GAAG,IAAI,CAAC,QAAQ,EAAE,GAAG,UAAU,EAAE;QAC7C,iBAAiB,EAAE,sBAAsB,CAAC,IAAI,CAAC,iBAAiB,EAAE,kBAAkB,CAAC;KACtF,CAAC;AACJ,CAAC;AAED,KAAK,UAAU,gBAAgB,CAAC,OAAqB,EAAE,GAAW,EAAE,IAAiB;IACnF,MAAM,OAAO,GAAG,IAAI,eAAe,EAAE,CAAC;IACtC,MAAM,KAAK,GAAG,UAAU,CAAC,GAAG,EAAE,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,4BAA4B,CAAC,CAAC;IAC9E,MAAM,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC,IAAI,CAAC,MAAM,EAAE,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,MAAM,CAAC;IAC7F,IAAI,CAAC;QACH,MAAM,QAAQ,GAAG,MAAM,OAAO,CAAC,GAAG,EAAE,EAAE,GAAG,IAAI,EAAE,MAAM,EAAE,QAAQ,EAAE,IAAI,CAAC,QAAQ,IAAI,QAAQ,EAAE,CAAC,CAAC;QAC9F,IAAI,CAAC;YACH,IAAI,CAAC,QAAQ,CAAC,EAAE;gBAAE,MAAM,IAAI,KAAK,CAAC,qCAAqC,QAAQ,CAAC,MAAM,EAAE,CAAC,CAAC;YAC1F,OAAO,MAAM,gBAAgB,CAAC,QAAQ,EAAE,EAAE,MAAM,EAAE,QAAQ,EAAE,0BAA0B,EAAE,CAAC,CAAC;QAC5F,CAAC;QAAC,OAAO,KAAK,EAAE,CAAC;YACf,MAAM,UAAU,CAAC,QAAQ,CAAC,CAAC;YAC3B,MAAM,KAAK,CAAC;QACd,CAAC;IACH,CAAC;YAAS,CAAC;QACT,YAAY,CAAC,KAAK,CAAC,CAAC;IACtB,CAAC;AACH,CAAC;AAED,SAAS,sBAAsB,CAC7B,IAA0C,EAC1C,UAAuE;IAEvE,MAAM,GAAG,GAA4C,EAAE,CAAC;IACxD,IAAI,IAAI;QACN,KAAK,MAAM,CAAC,SAAS,EAAE,QAAQ,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,IAAI,CAAC;YACtD,GAAG,CAAC,SAAS,CAAC,GAAG,EAAE,GAAI,QAAoC,EAAE,CAAC;IAClE,KAAK,MAAM,CAAC,SAAS,EAAE,QAAQ,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,UAAU,CAAC;QAC5D,GAAG,CAAC,SAAS,CAAC,GAAG,EAAE,GAAG,CAAC,GAAG,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,EAAE,GAAG,QAAQ,EAAE,CAAC;IAC9D,OAAO,GAAG,CAAC;AACb,CAAC;AAED,SAAS,aAAa,CAAC,MAAc;IACnC,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,IAAI,GAAG,CAAC,MAAM,EAAE,eAAe,CAAC,CAAC;QAC7C,IAAI,GAAG,CAAC,QAAQ,KAAK,QAAQ;YAAE,OAAO,SAAS,CAAC;QAChD,IAAI,GAAG,CAAC,QAAQ,KAAK,OAAO,IAAI,GAAG,CAAC,QAAQ,KAAK,eAAe;YAAE,OAAO,SAAS,CAAC;QACnF,OAAO,GAAG,CAAC,QAAQ,EAAE,CAAC;IACxB,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,SAAS,CAAC;IACnB,CAAC;AACH,CAAC;AAED,SAAS,cAAc,CAAC,MAAc;IACpC,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,MAAM,CAAC,CAAC,QAAQ,CAAC;IACtC,IAAI,iDAAiD,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,CAAC,CAAC;IAC3E,IAAI,qBAAqB,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,CAAC,CAAC;IAC/C,OAAO,CAAC,CAAC;AACX,CAAC"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"csv-writer.d.ts","sourceRoot":"","sources":["../../src/output/csv-writer.ts"],"names":[],"mappings":"
|
|
1
|
+
{"version":3,"file":"csv-writer.d.ts","sourceRoot":"","sources":["../../src/output/csv-writer.ts"],"names":[],"mappings":"AAMA,wBAAsB,QAAQ,CAC5B,IAAI,EAAE,MAAM,EACZ,IAAI,EAAE,SAAS,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,EACxC,MAAM,UAAQ,GACb,OAAO,CAAC,IAAI,CAAC,CAwBf"}
|
|
@@ -1,18 +1,40 @@
|
|
|
1
|
-
import { access, mkdir, readFile, writeFile } from "node:fs/promises";
|
|
1
|
+
import { access, appendFile, mkdir, readFile, rename, writeFile } from "node:fs/promises";
|
|
2
2
|
import { dirname } from "node:path";
|
|
3
|
+
import { randomUUID } from "node:crypto";
|
|
4
|
+
import { EngineError } from "../domain/errors.js";
|
|
5
|
+
import { withDestinationLock } from "./lock.js";
|
|
3
6
|
export async function writeCsv(path, rows, append = false) {
|
|
4
7
|
if (rows.length === 0)
|
|
5
8
|
return;
|
|
6
|
-
|
|
7
|
-
|
|
8
|
-
|
|
9
|
-
|
|
10
|
-
|
|
11
|
-
|
|
12
|
-
|
|
13
|
-
|
|
9
|
+
await withDestinationLock(path, async () => {
|
|
10
|
+
const incomingKeys = rows.flatMap((row) => Object.keys(row));
|
|
11
|
+
if (append && (await exists(path))) {
|
|
12
|
+
const existing = await readFile(path, "utf8");
|
|
13
|
+
const parsedExisting = existing ? parseCsv(existing) : [];
|
|
14
|
+
const currentHeaders = parsedExisting[0] ?? [];
|
|
15
|
+
if (currentHeaders.length === 0)
|
|
16
|
+
throw new EngineError(`Refusing to append to a CSV file without a header: ${path}`);
|
|
17
|
+
const nextHeaders = mergeHeaders(currentHeaders, incomingKeys);
|
|
18
|
+
if (!sameHeaders(currentHeaders, nextHeaders))
|
|
19
|
+
throw new EngineError(`CSV schema changed for ${path}; write a new snapshot instead of appending mixed columns.`);
|
|
20
|
+
const body = rows
|
|
21
|
+
.map((row) => currentHeaders.map((header) => escapeCsv(flattenValue(row[header]))).join(","))
|
|
22
|
+
.join("\n");
|
|
23
|
+
await appendFile(path, `${body}\n`, "utf8");
|
|
24
|
+
return;
|
|
25
|
+
}
|
|
26
|
+
const headers = mergeHeaders([], incomingKeys);
|
|
27
|
+
await atomicWrite(path, renderCsv(headers, rows));
|
|
28
|
+
});
|
|
29
|
+
}
|
|
30
|
+
function sameHeaders(left, right) {
|
|
31
|
+
return left.length === right.length && left.every((header, index) => header === right[index]);
|
|
32
|
+
}
|
|
33
|
+
async function atomicWrite(path, content) {
|
|
14
34
|
await mkdir(dirname(path), { recursive: true });
|
|
15
|
-
|
|
35
|
+
const temporary = `${path}.${randomUUID()}.tmp`;
|
|
36
|
+
await writeFile(temporary, content, "utf8");
|
|
37
|
+
await rename(temporary, path);
|
|
16
38
|
}
|
|
17
39
|
function renderCsv(headers, rows) {
|
|
18
40
|
const body = rows.map((row) => headers.map((header) => escapeCsv(flattenValue(row[header]))).join(","));
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"csv-writer.js","sourceRoot":"","sources":["../../src/output/csv-writer.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,EAAE,KAAK,EAAE,QAAQ,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;
|
|
1
|
+
{"version":3,"file":"csv-writer.js","sourceRoot":"","sources":["../../src/output/csv-writer.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,EAAE,UAAU,EAAE,KAAK,EAAE,QAAQ,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AAC1F,OAAO,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AACpC,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AACzC,OAAO,EAAE,WAAW,EAAE,MAAM,qBAAqB,CAAC;AAClD,OAAO,EAAE,mBAAmB,EAAE,MAAM,WAAW,CAAC;AAEhD,MAAM,CAAC,KAAK,UAAU,QAAQ,CAC5B,IAAY,EACZ,IAAwC,EACxC,MAAM,GAAG,KAAK;IAEd,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO;IAC9B,MAAM,mBAAmB,CAAC,IAAI,EAAE,KAAK,IAAI,EAAE;QACzC,MAAM,YAAY,GAAG,IAAI,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;QAC7D,IAAI,MAAM,IAAI,CAAC,MAAM,MAAM,CAAC,IAAI,CAAC,CAAC,EAAE,CAAC;YACnC,MAAM,QAAQ,GAAG,MAAM,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC,CAAC;YAC9C,MAAM,cAAc,GAAG,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;YAC1D,MAAM,cAAc,GAAG,cAAc,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC;YAC/C,IAAI,cAAc,CAAC,MAAM,KAAK,CAAC;gBAC7B,MAAM,IAAI,WAAW,CAAC,sDAAsD,IAAI,EAAE,CAAC,CAAC;YACtF,MAAM,WAAW,GAAG,YAAY,CAAC,cAAc,EAAE,YAAY,CAAC,CAAC;YAC/D,IAAI,CAAC,WAAW,CAAC,cAAc,EAAE,WAAW,CAAC;gBAC3C,MAAM,IAAI,WAAW,CACnB,0BAA0B,IAAI,4DAA4D,CAC3F,CAAC;YACJ,MAAM,IAAI,GAAG,IAAI;iBACd,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,SAAS,CAAC,YAAY,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;iBAC5F,IAAI,CAAC,IAAI,CAAC,CAAC;YACd,MAAM,UAAU,CAAC,IAAI,EAAE,GAAG,IAAI,IAAI,EAAE,MAAM,CAAC,CAAC;YAC5C,OAAO;QACT,CAAC;QACD,MAAM,OAAO,GAAG,YAAY,CAAC,EAAE,EAAE,YAAY,CAAC,CAAC;QAC/C,MAAM,WAAW,CAAC,IAAI,EAAE,SAAS,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC,CAAC;IACpD,CAAC,CAAC,CAAC;AACL,CAAC;AAED,SAAS,WAAW,CAAC,IAAuB,EAAE,KAAwB;IACpE,OAAO,IAAI,CAAC,MAAM,KAAK,KAAK,CAAC,MAAM,IAAI,IAAI,CAAC,KAAK,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,CAAC,MAAM,KAAK,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC;AAChG,CAAC;AAED,KAAK,UAAU,WAAW,CAAC,IAAY,EAAE,OAAe;IACtD,MAAM,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;IAChD,MAAM,SAAS,GAAG,GAAG,IAAI,IAAI,UAAU,EAAE,MAAM,CAAC;IAChD,MAAM,SAAS,CAAC,SAAS,EAAE,OAAO,EAAE,MAAM,CAAC,CAAC;IAC5C,MAAM,MAAM,CAAC,SAAS,EAAE,IAAI,CAAC,CAAC;AAChC,CAAC;AAED,SAAS,SAAS,CAAC,OAA0B,EAAE,IAAwC;IACrF,MAAM,IAAI,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,SAAS,CAAC,YAAY,CAAC,GAAG,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;IACxG,OAAO,GAAG,OAAO,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC;AACrE,CAAC;AAED,SAAS,YAAY,CAAC,QAA2B,EAAE,QAA2B;IAC5E,MAAM,SAAS,GAAG;QAChB,SAAS;QACT,WAAW;QACX,UAAU;QACV,MAAM;QACN,MAAM;QACN,OAAO;QACP,UAAU;QACV,UAAU;QACV,UAAU;KACX,CAAC;IACF,MAAM,GAAG,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,QAAQ,EAAE,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;IACrD,OAAO,CAAC,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,EAAE,GAAG,GAAG,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,SAAS,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC;AAC7G,CAAC;AAED,SAAS,QAAQ,CAAC,KAAa;IAC7B,MAAM,IAAI,GAAe,EAAE,CAAC;IAC5B,IAAI,GAAG,GAAa,EAAE,CAAC;IACvB,IAAI,KAAK,GAAG,EAAE,CAAC;IACf,IAAI,MAAM,GAAG,KAAK,CAAC;IACnB,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,KAAK,CAAC,MAAM,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QACrD,MAAM,SAAS,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC;QAC/B,MAAM,IAAI,GAAG,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;QAC9B,IAAI,SAAS,KAAK,GAAG,EAAE,CAAC;YACtB,IAAI,MAAM,IAAI,IAAI,KAAK,GAAG,EAAE,CAAC;gBAC3B,KAAK,IAAI,GAAG,CAAC;gBACb,KAAK,IAAI,CAAC,CAAC;YACb,CAAC;;gBAAM,MAAM,GAAG,CAAC,MAAM,CAAC;QAC1B,CAAC;aAAM,IAAI,SAAS,KAAK,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;YACxC,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YAChB,KAAK,GAAG,EAAE,CAAC;QACb,CAAC;aAAM,IAAI,CAAC,SAAS,KAAK,IAAI,IAAI,SAAS,KAAK,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,CAAC;YACjE,IAAI,SAAS,KAAK,IAAI,IAAI,IAAI,KAAK,IAAI;gBAAE,KAAK,IAAI,CAAC,CAAC;YACpD,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YAChB,KAAK,GAAG,EAAE,CAAC;YACX,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC;gBAAE,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACxD,GAAG,GAAG,EAAE,CAAC;QACX,CAAC;aAAM,CAAC;YACN,KAAK,IAAI,SAAS,CAAC;QACrB,CAAC;IACH,CAAC;IACD,IAAI,KAAK,IAAI,GAAG,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC5B,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAChB,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IACjB,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,SAAS,YAAY,CAAC,KAAc;IAClC,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,KAAK,IAAI;QAAE,OAAO,EAAE,CAAC;IACrD,IAAI,OAAO,KAAK,KAAK,QAAQ;QAAE,OAAO,KAAK,CAAC;IAC5C,OAAO,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC;AAC/B,CAAC;AAED,SAAS,SAAS,CAAC,KAAa;IAC9B,OAAO,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,UAAU,CAAC,GAAG,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC;AAC9E,CAAC;AAED,KAAK,UAAU,MAAM,CAAC,IAAY;IAChC,IAAI,CAAC;QACH,MAAM,MAAM,CAAC,IAAI,CAAC,CAAC;QACnB,OAAO,IAAI,CAAC;IACd,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,KAAK,CAAC;IACf,CAAC;AACH,CAAC"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"json-writer.d.ts","sourceRoot":"","sources":["../../src/output/json-writer.ts"],"names":[],"mappings":"
|
|
1
|
+
{"version":3,"file":"json-writer.d.ts","sourceRoot":"","sources":["../../src/output/json-writer.ts"],"names":[],"mappings":"AAMA,wBAAsB,SAAS,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,SAAS,OAAO,EAAE,EAAE,MAAM,UAAQ,GAAG,OAAO,CAAC,IAAI,CAAC,CAiBrG;AAED,wBAAsB,WAAW,CAAC,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,SAAS,OAAO,EAAE,EAAE,MAAM,UAAQ,GAAG,OAAO,CAAC,IAAI,CAAC,CAWvG"}
|
|
@@ -1,21 +1,46 @@
|
|
|
1
|
-
import { access, mkdir, readFile, writeFile } from "node:fs/promises";
|
|
1
|
+
import { access, appendFile, mkdir, readFile, rename, writeFile } from "node:fs/promises";
|
|
2
2
|
import { dirname } from "node:path";
|
|
3
|
+
import { randomUUID } from "node:crypto";
|
|
4
|
+
import { EngineError } from "../domain/errors.js";
|
|
5
|
+
import { withDestinationLock } from "./lock.js";
|
|
3
6
|
export async function writeJson(path, rows, append = false) {
|
|
4
7
|
if (rows.length === 0)
|
|
5
8
|
return;
|
|
6
|
-
|
|
7
|
-
|
|
8
|
-
|
|
9
|
-
|
|
10
|
-
|
|
11
|
-
|
|
9
|
+
await withDestinationLock(path, async () => {
|
|
10
|
+
let output = [...rows];
|
|
11
|
+
if (append && (await exists(path))) {
|
|
12
|
+
let current;
|
|
13
|
+
try {
|
|
14
|
+
current = JSON.parse(await readFile(path, "utf8"));
|
|
15
|
+
}
|
|
16
|
+
catch {
|
|
17
|
+
throw new EngineError(`Refusing to append to malformed JSON file: ${path}`);
|
|
18
|
+
}
|
|
19
|
+
if (!Array.isArray(current))
|
|
20
|
+
throw new EngineError(`Refusing to append to a JSON file that is not an array: ${path}`);
|
|
21
|
+
output = [...current, ...rows];
|
|
12
22
|
}
|
|
13
|
-
|
|
14
|
-
|
|
23
|
+
await atomicWrite(path, `${JSON.stringify(output, null, 2)}\n`);
|
|
24
|
+
});
|
|
25
|
+
}
|
|
26
|
+
export async function writeNdjson(path, rows, append = false) {
|
|
27
|
+
if (rows.length === 0)
|
|
28
|
+
return;
|
|
29
|
+
await withDestinationLock(path, async () => {
|
|
30
|
+
const chunk = `${rows.map((row) => JSON.stringify(row)).join("\n")}\n`;
|
|
31
|
+
if (!append || !(await exists(path))) {
|
|
32
|
+
await atomicWrite(path, chunk);
|
|
33
|
+
return;
|
|
15
34
|
}
|
|
16
|
-
|
|
35
|
+
await mkdir(dirname(path), { recursive: true });
|
|
36
|
+
await appendFile(path, chunk, "utf8");
|
|
37
|
+
});
|
|
38
|
+
}
|
|
39
|
+
async function atomicWrite(path, content) {
|
|
17
40
|
await mkdir(dirname(path), { recursive: true });
|
|
18
|
-
|
|
41
|
+
const temporary = `${path}.${randomUUID()}.tmp`;
|
|
42
|
+
await writeFile(temporary, content, "utf8");
|
|
43
|
+
await rename(temporary, path);
|
|
19
44
|
}
|
|
20
45
|
async function exists(path) {
|
|
21
46
|
try {
|