mirror of
https://github.com/renovatebot/renovate.git
synced 2026-11-01 02:00:10 +00:00
3412de84ce
* refactor(lint): enable oxlint prefer-named-capture-group at error Numbered capture groups (m[1], $1) break silently when a regex is edited and group order shifts. Named groups (?<name>...) make intent explicit and accesses (m.groups.name) resilient to reordering. This enables the built-in eslint rule prefer-named-capture-group, surfacing 321 pre-existing unnamed-group violations across ~121 files to be burned down incrementally in follow-up commits. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017RxHFNsSworLcv23xXk9Hk * refactor(versioning): name capture groups for prefer-named-capture-group Burns down all prefer-named-capture-group violations in lib/modules/versioning/. Unused capture groups (only reachable via .test() or whole-match) become non-capturing (?:...); groups whose value is read become named (?<name>...), with accesses rewritten to match.groups!.<name> (or destructured from .groups!) instead of positional match[n]/$n so a future edit that reorders groups cannot silently break the reader. Two cases needed extra care to preserve behavior: - semver/common.ts and poetry/patterns.ts split() on a regex whose capture group is relied upon by String.prototype.split to keep the separator in the output array; the group is kept capturing and named, not converted to non-capturing. - hashicorp/convertor.ts's semverRegex is only ever embedded via .source into a named `(?<version>...)` group elsewhere and never matched directly, so its internal groups are safe to make non-capturing. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017RxHFNsSworLcv23xXk9Hk * refactor(datasource): name capture groups for prefer-named-capture-group Burns down all prefer-named-capture-group violations under lib/modules/datasource/. Same treatment as the versioning module: unused groups become non-capturing, used groups get names, and positional match[n] reads move to match.groups!.<name>. docker/ecr.ts and pypi/common.ts keep their positional destructuring (`const [, region] = ...`, `!!m &&`) working unchanged since a named capture group is still addressable by its numeric index in JavaScript — only pypi/common.ts's access was rewritten to m.groups!.owner for clarity since it read a single group directly. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017RxHFNsSworLcv23xXk9Hk * refactor(platform): name capture groups for prefer-named-capture-group Burns down all prefer-named-capture-group violations under lib/modules/platform/. Uses the repo's established idiom of `match.groups?.<name>` after an optional-chained exec()/match() (see bitbucket/index.ts hostnameWithoutApiPrefix, gerrit/utils.ts extractSourceBranch) and `$<name>` in String.replace replacement strings. Unused groups (markdown/HTML massaging regexes matched only to be stripped) become non-capturing. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017RxHFNsSworLcv23xXk9Hk * refactor(manager): name capture groups in ansible-galaxy for prefer-named-capture-group Burns down prefer-named-capture-group violations in lib/modules/manager/ansible-galaxy/. newBlockRegEx and blockLineRegEx dropped their unused outer wrapping group (only the inner key/value groups were ever read) and named the key/value groups; nameMatchRegex's remaining unused groups (protocol/user-info/port alternatives) become non-capturing since only source/hostname/depName/version are read via .groups. Accesses in collections.ts and roles.ts move from lineMatch[2]/lineMatch[3] to lineMatch.groups!.key/value. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017RxHFNsSworLcv23xXk9Hk * refactor(manager): name capture groups for prefer-named-capture-group (batch 1) Burns down prefer-named-capture-group violations across the first alphabetical slice of lib/modules/manager/ (ansible through mix). Used groups get names with accesses rewritten to match.groups!.<name> (or $<name> in replacement strings); groups whose value is never read become non-capturing (?:...). Two String.split() call sites (git-submodules/extract.ts, meteor/extract.ts) keep the established "named group still readable positionally from the split() result array" idiom instead of restructuring the destructuring. bazel/rules/go.ts and buildpacks/extract.ts dockerRef verified byte-for-byte equivalent before/after via a quick node script since several nested groups collapsed to non-capturing at once. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017RxHFNsSworLcv23xXk9Hk * refactor(manager): name capture groups for prefer-named-capture-group (batch 2) Burns down prefer-named-capture-group violations across the remaining alphabetical slice of lib/modules/manager/ (npm through terragrunt). Same treatment as batch 1: used groups get names with accesses moved to match.groups!.<name> or $<name> replacement strings; unused groups (wrapping alternations, quantifier repetition groups only consumed via .test(), or values never read from a match) become non-capturing. terraform/extractors/others/modules.ts and terragrunt/modules.ts share near-identical URL-parsing regexes (github/bitbucket/git/azure-devops ref matchers) — the same set of always-unused `depth=`/`&depth=` query-string groups is converted to non-capturing in both files. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017RxHFNsSworLcv23xXk9Hk * refactor(config): name capture groups for prefer-named-capture-group Burns down prefer-named-capture-group violations in lib/config/. Used groups get names with replacement strings switched from $1/$2 to $<name>; the unused separator group in schedule-migration.ts's afterBeforeRegex becomes non-capturing. presets/parse.ts keeps assigning through the already-declared `repo` variable rather than array-destructuring the exec() result, matching this file's existing style for the other named-group regexes just above it. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017RxHFNsSworLcv23xXk9Hk * refactor(util): name capture groups for prefer-named-capture-group Burns down prefer-named-capture-group violations across lib/util/ and lib/logger/. Same treatment as previous batches: named groups for values that are read (via $<name> replacement strings, .groups!.<name>, or destructured from .groups), non-capturing groups where nothing reads the captured value (e.g. semantic.ts's angular-commit .test()-only regex). www-authenticate.ts's tokenizer entries each get their own named group but keep accessing the match generically via match[1], since a positionally-indexed named group is still readable by index. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017RxHFNsSworLcv23xXk9Hk * refactor(workers): name capture groups for prefer-named-capture-group Burns down prefer-named-capture-group violations in lib/workers/. Used groups get names with $<name> replacement strings; unused wrapper/alternation groups (e.g. semantic-commit-message.ts's optional scope/issue wrappers, release-notes.ts's date-heading and horizontal-rule matchers that are only probed with .search()/replaced wholesale) become non-capturing. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017RxHFNsSworLcv23xXk9Hk * refactor(test,tools): name capture groups for prefer-named-capture-group Finishes the prefer-named-capture-group burndown: test/, tools/, and vitest.config.mts. Groups whose value is read get names ($<name> replacement strings or .groups.<name> access); groups only used via .test()/.search() or that feed a .match(/…/g) call (whose return value is the array of whole matches, not captured groups) become non-capturing. This is the last remaining slice of pre-existing violations, so the rule (already enabled at error repo-wide) now passes with zero findings across the whole codebase. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_017RxHFNsSworLcv23xXk9Hk * fix prettier failure * refactor(lint): use a named capture group in `v8-ignore-no-count` New rule from main trips the newly enabled `prefer-named-capture-group` oxlint rule. --------- Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
459 lines
17 KiB
TypeScript
459 lines
17 KiB
TypeScript
import { isNonEmptyString } from '@sindresorhus/is';
|
|
import changelogFilenameRegex from 'changelog-filename-regex';
|
|
import { logger } from '../../../logger/index.ts';
|
|
import { ExternalHostError } from '../../../types/errors/external-host-error.ts';
|
|
import { coerceArray, deduplicateArray } from '../../../util/array.ts';
|
|
import { getEnv } from '../../../util/env.ts';
|
|
import { parse } from '../../../util/html.ts';
|
|
import { HttpError } from '../../../util/http/index.ts';
|
|
import type {
|
|
HttpResponse,
|
|
OutgoingHttpHeaders,
|
|
} from '../../../util/http/types.ts';
|
|
import { regEx } from '../../../util/regex.ts';
|
|
import { Json } from '../../../util/schema-utils/index.ts';
|
|
import type { Timestamp } from '../../../util/timestamp.ts';
|
|
import { asTimestamp } from '../../../util/timestamp.ts';
|
|
import { ensureTrailingSlash, parseUrl } from '../../../util/url.ts';
|
|
import * as pep440 from '../../versioning/pep440/index.ts';
|
|
import { Datasource } from '../datasource.ts';
|
|
import type { GetReleasesConfig, Release, ReleaseResult } from '../types.ts';
|
|
import { getGoogleAuthToken } from '../util.ts';
|
|
import { isGitHubRepo, normalizePythonDepName } from './common.ts';
|
|
import type { PypiRelease } from './schema.ts';
|
|
import { PypiResponse, PypiSimpleResponse } from './schema.ts';
|
|
import type { Releases } from './types.ts';
|
|
|
|
export class PypiDatasource extends Datasource {
|
|
static readonly id = 'pypi';
|
|
|
|
constructor() {
|
|
super(PypiDatasource.id);
|
|
}
|
|
|
|
override readonly caching = true;
|
|
|
|
override readonly customRegistrySupport = true;
|
|
|
|
static readonly defaultURL =
|
|
getEnv().PIP_INDEX_URL ?? 'https://pypi.org/pypi/';
|
|
override readonly defaultRegistryUrls = [PypiDatasource.defaultURL];
|
|
|
|
override readonly defaultVersioning = pep440.id;
|
|
|
|
override readonly registryStrategy = 'merge';
|
|
|
|
override readonly releaseTimestampSupport = true;
|
|
override readonly releaseTimestampNote =
|
|
'The release timestamp is determined from the earliest `upload_time` field of the files of a version. When using the Simple API, timestamps are available if the server supports the JSON-based Simple API (PEP 691).';
|
|
override readonly sourceUrlSupport = 'release';
|
|
override readonly sourceUrlNote =
|
|
'The source URL is determined from the `homepage` field if it is a github repository, else we use the `project_urls` field.';
|
|
|
|
async getReleases({
|
|
packageName,
|
|
registryUrl,
|
|
}: GetReleasesConfig): Promise<ReleaseResult | null> {
|
|
let dependency: ReleaseResult | null = null;
|
|
// TODO: null check (#22198)
|
|
const hostUrl = ensureTrailingSlash(
|
|
registryUrl!.replace('https://pypi.org/simple', 'https://pypi.org/pypi'),
|
|
);
|
|
const normalizedLookupName = normalizePythonDepName(packageName);
|
|
|
|
// not all simple indexes use this identifier, but most do
|
|
if (hostUrl.endsWith('/simple/') || hostUrl.endsWith('/+simple/')) {
|
|
logger.trace(
|
|
{ packageName, hostUrl },
|
|
'Looking up pypi simple dependency',
|
|
);
|
|
dependency = await this.getSimpleDependency(
|
|
normalizedLookupName,
|
|
hostUrl,
|
|
);
|
|
} else {
|
|
logger.trace({ packageName, hostUrl }, 'Looking up pypi api dependency');
|
|
try {
|
|
// we need to resolve early here so we can catch any 404s and fallback to a simple lookup
|
|
dependency = await this.getDependency(normalizedLookupName, hostUrl);
|
|
} catch (err) {
|
|
// error contacting json-style api -- attempt to fallback to a simple-style api
|
|
logger.trace(
|
|
{ packageName, hostUrl, err },
|
|
'Looking up pypi simple dependency via fallback',
|
|
);
|
|
dependency = await this.getSimpleDependency(
|
|
normalizedLookupName,
|
|
hostUrl,
|
|
);
|
|
}
|
|
}
|
|
return dependency;
|
|
}
|
|
|
|
private sanitizeLookupUrl(lookupUrl: string, parsedUrl: URL): string {
|
|
if (!parsedUrl.username && !parsedUrl.password) {
|
|
return lookupUrl;
|
|
}
|
|
|
|
parsedUrl.username = '';
|
|
parsedUrl.password = '';
|
|
return parsedUrl.toString();
|
|
}
|
|
|
|
private async getAuthHeaders(
|
|
lookupUrl: string,
|
|
): Promise<{ headers: OutgoingHttpHeaders; lookupUrl: string }> {
|
|
const parsedUrl = parseUrl(lookupUrl);
|
|
// v8 ignore if -- TODO: refactor to cover this branch through public behavior again
|
|
if (!parsedUrl) {
|
|
logger.once.debug({ lookupUrl }, 'Failed to parse URL');
|
|
return { headers: {}, lookupUrl };
|
|
}
|
|
if (parsedUrl.hostname.endsWith('.pkg.dev')) {
|
|
const auth = await getGoogleAuthToken();
|
|
if (auth) {
|
|
const sanitizedLookupUrl = this.sanitizeLookupUrl(lookupUrl, parsedUrl);
|
|
return {
|
|
headers: { authorization: `Basic ${auth}` },
|
|
lookupUrl: sanitizedLookupUrl,
|
|
};
|
|
}
|
|
logger.once.debug({ lookupUrl }, 'Could not get Google access token');
|
|
return { headers: {}, lookupUrl };
|
|
}
|
|
return { headers: {}, lookupUrl };
|
|
}
|
|
|
|
private async getDependency(
|
|
packageName: string,
|
|
hostUrl: string,
|
|
): Promise<ReleaseResult | null> {
|
|
const lookupUrl = new URL(
|
|
`${normalizePythonDepName(packageName)}/json`,
|
|
hostUrl,
|
|
).href;
|
|
const dependency: ReleaseResult = { releases: [] };
|
|
logger.trace({ lookupUrl }, 'Pypi api got lookup');
|
|
const { headers, lookupUrl: sanitizedUrl } =
|
|
await this.getAuthHeaders(lookupUrl);
|
|
const rep = await this.http.getJson(
|
|
sanitizedUrl,
|
|
{ headers },
|
|
PypiResponse,
|
|
);
|
|
const dep = rep?.body;
|
|
if (rep.authorization) {
|
|
dependency.isPrivate = true;
|
|
}
|
|
logger.trace({ lookupUrl }, 'Got pypi api result');
|
|
|
|
if (dep.info?.home_page) {
|
|
dependency.homepage = dep.info.home_page;
|
|
if (isGitHubRepo(dep.info.home_page)) {
|
|
dependency.sourceUrl = dep.info.home_page.replace(
|
|
'http://',
|
|
'https://',
|
|
);
|
|
}
|
|
}
|
|
|
|
if (dep.info?.project_urls) {
|
|
for (const [name, projectUrl] of Object.entries(dep.info.project_urls)) {
|
|
const lower = name.toLowerCase();
|
|
|
|
if (
|
|
projectUrl &&
|
|
!dependency.sourceUrl &&
|
|
(lower.startsWith('repo') ||
|
|
lower === 'code' ||
|
|
lower === 'source' ||
|
|
isGitHubRepo(projectUrl))
|
|
) {
|
|
dependency.sourceUrl = projectUrl;
|
|
}
|
|
|
|
if (
|
|
!dependency.changelogUrl &&
|
|
([
|
|
'changelog',
|
|
'change log',
|
|
'changes',
|
|
'release notes',
|
|
'news',
|
|
"what's new",
|
|
].includes(lower) ||
|
|
changelogFilenameRegex.exec(lower))
|
|
) {
|
|
// from https://github.com/pypa/warehouse/blob/418c7511dc367fb410c71be139545d0134ccb0df/warehouse/templates/packaging/detail.html#L24
|
|
dependency.changelogUrl = projectUrl;
|
|
}
|
|
}
|
|
}
|
|
|
|
if (dep.releases) {
|
|
dependency.releases = PypiDatasource.toReleases(dep.releases);
|
|
}
|
|
return dependency;
|
|
}
|
|
|
|
private static getEarliestTimestamp(
|
|
releases: PypiRelease[],
|
|
): Timestamp | null {
|
|
let earliest: Timestamp | null = null;
|
|
for (const { upload_time } of releases) {
|
|
const timestamp = asTimestamp(upload_time);
|
|
// `asTimestamp` normalizes to UTC ISO 8601, so comparing the strings compares the instants
|
|
if (timestamp && (!earliest || timestamp < earliest)) {
|
|
earliest = timestamp;
|
|
}
|
|
}
|
|
return earliest;
|
|
}
|
|
|
|
private static extractVersionFromLinkText(
|
|
text: string,
|
|
packageName: string,
|
|
): string | null {
|
|
// source packages
|
|
const lcText = text.toLowerCase();
|
|
const normalizedSrcText = normalizePythonDepName(text);
|
|
const srcPrefix = `${packageName}-`;
|
|
|
|
// source distribution format: `{name}-{version}.tar.gz` (https://packaging.python.org/en/latest/specifications/source-distribution-format/#source-distribution-file-name)
|
|
// binary distribution: `{distribution}-{version}(-{build tag})?-{python tag}-{abi tag}-{platform tag}.whl` (https://packaging.python.org/en/latest/specifications/binary-distribution-format/#file-name-convention)
|
|
// officially both `name` and `distribution` should be normalized and then the - replaced with _, but in reality this is not the case
|
|
// We therefore normalize the name we have (replacing `_-.` with -) and then check if the text starts with the normalized name
|
|
|
|
if (!normalizedSrcText.startsWith(srcPrefix)) {
|
|
return null;
|
|
}
|
|
|
|
// strip off the prefix using the prefix length as we may have normalized the srcPrefix/packageName
|
|
// We assume that neither the version nor the suffix contains multiple `-` like `0.1.2---rc1.tar.gz`
|
|
// and use the difference in length to strip off the prefix in case the name contains double `--` characters
|
|
const normalizedLengthDiff = lcText.length - normalizedSrcText.length;
|
|
const res = lcText.slice(srcPrefix.length + normalizedLengthDiff);
|
|
|
|
// source distribution
|
|
const srcSuffixes = ['.tar.gz', '.tar.bz2', '.tar.xz', '.zip', '.tgz'];
|
|
const srcSuffix = srcSuffixes.find((suffix) => lcText.endsWith(suffix));
|
|
if (srcSuffix) {
|
|
// strip off the suffix using character length
|
|
return res.slice(0, -srcSuffix.length);
|
|
}
|
|
|
|
// binary distribution
|
|
// for binary distributions the version is the first part after the removed distribution name
|
|
const wheelSuffix = '.whl';
|
|
if (lcText.endsWith(wheelSuffix) && lcText.split('-').length > 2) {
|
|
return res.split('-')[0];
|
|
}
|
|
return null;
|
|
}
|
|
|
|
private static cleanSimpleHtml(html: string): string {
|
|
return (
|
|
html
|
|
.replace(regEx(/<\/?pre>/), '')
|
|
// Certain simple repositories like artifactory don't escape > and <
|
|
.replace(
|
|
regEx(/data-requires-python="(?<before>[^"]*?)>(?<after>[^"]*?)"/g),
|
|
'data-requires-python="$<before>>$<after>"',
|
|
)
|
|
.replace(
|
|
regEx(/data-requires-python="(?<before>[^"]*?)<(?<after>[^"]*?)"/g),
|
|
'data-requires-python="$<before><$<after>"',
|
|
)
|
|
);
|
|
}
|
|
|
|
private static getSimpleReleasesFromHtml(
|
|
html: string,
|
|
packageName: string,
|
|
): Releases {
|
|
const root = parse(PypiDatasource.cleanSimpleHtml(html));
|
|
const links = root.querySelectorAll('a');
|
|
const releases: Releases = {};
|
|
for (const link of Array.from(links)) {
|
|
const version = PypiDatasource.extractVersionFromLinkText(
|
|
link.text?.trim(),
|
|
packageName,
|
|
);
|
|
if (version) {
|
|
const release: PypiRelease = {
|
|
yanked: link.hasAttribute('data-yanked'),
|
|
};
|
|
const requiresPython = link.getAttribute('data-requires-python');
|
|
if (requiresPython) {
|
|
release.requires_python = requiresPython;
|
|
}
|
|
(releases[version] ??= []).push(release);
|
|
}
|
|
}
|
|
return releases;
|
|
}
|
|
|
|
private static getSimpleReleasesFromJson(
|
|
json: string,
|
|
packageName: string,
|
|
): Releases | null {
|
|
const releases: Releases = {};
|
|
const parsed = Json.pipe(PypiSimpleResponse).safeParse(json);
|
|
if (!parsed.success) {
|
|
logger.once.warn(
|
|
{ packageName, err: parsed.error },
|
|
'Failed to parse JSON-based Simple API response',
|
|
);
|
|
// Distinguish a malformed response from a package with genuinely no files, so the caller doesn't mistake an error for an empty result.
|
|
return null;
|
|
}
|
|
for (const file of parsed.data.files) {
|
|
const version = PypiDatasource.extractVersionFromLinkText(
|
|
file.filename,
|
|
packageName,
|
|
);
|
|
if (version) {
|
|
(releases[version] ??= []).push(file);
|
|
}
|
|
}
|
|
return releases;
|
|
}
|
|
|
|
private async getSimpleDependency(
|
|
packageName: string,
|
|
hostUrl: string,
|
|
): Promise<ReleaseResult | null> {
|
|
const lookupUrl = new URL(
|
|
ensureTrailingSlash(normalizePythonDepName(packageName)),
|
|
hostUrl,
|
|
).href;
|
|
const dependency: ReleaseResult = { releases: [] };
|
|
const { headers: authHeaders, lookupUrl: sanitizedUrl } =
|
|
await this.getAuthHeaders(lookupUrl);
|
|
const headers: OutgoingHttpHeaders = {
|
|
...authHeaders,
|
|
// Request the JSON serialization (PEP 691), falling back to the legacy HTML serialization (PEP 503) via content negotiation.
|
|
// https://github.com/pypa/pip/blob/6b0011b49a068c62f65389bf4cea7af5d28cb002/src/pip/_internal/index/collector.py#L128-L134
|
|
accept:
|
|
'application/vnd.pypi.simple.v1+json, application/vnd.pypi.simple.v1+html; q=0.1, text/html; q=0.01',
|
|
};
|
|
let response: HttpResponse;
|
|
try {
|
|
response = await this.http.getText(sanitizedUrl, { headers });
|
|
} catch (err) {
|
|
// An `abortOnError` host rule makes `Http` wrap the original error, so unwrap it before looking at the status code.
|
|
const httpErr = err instanceof ExternalHostError ? err.err : err;
|
|
const statusCode =
|
|
httpErr instanceof HttpError ? httpErr.response?.statusCode : undefined;
|
|
// A registry which cannot serve any of the negotiated types may pick a content type of its own, answer `406`, or answer `300`; the spec mandates none of them, but `406` is the common choice.
|
|
// https://packaging.python.org/en/latest/specifications/simple-repository-api/#version-format-selection
|
|
// Retry once without the negotiated `accept` header to preserve the pre-PEP-691 behaviour of a plain request.
|
|
if (statusCode !== 406) {
|
|
throw err;
|
|
}
|
|
|
|
logger.trace(
|
|
{ packageName, hostUrl, statusCode },
|
|
'Registry rejected negotiated Accept header, retrying without it',
|
|
);
|
|
response = await this.http.getText(sanitizedUrl, {
|
|
headers: authHeaders,
|
|
});
|
|
}
|
|
const dep = response?.body;
|
|
if (!dep) {
|
|
logger.trace({ dependency: packageName }, 'pip package not found');
|
|
return null;
|
|
}
|
|
if (response.authorization) {
|
|
dependency.isPrivate = true;
|
|
}
|
|
|
|
// Dispatch on the response content-type: JSON serialization (PEP 691) or the legacy HTML serialization (PEP 503). Matched loosely (any `json` media type) rather than the exact vendor type, as some registries/proxies relabel or strip the vendor-specific media type while still returning JSON.
|
|
const contentType = response.headers['content-type'];
|
|
const isJson = !!contentType && contentType.includes('json');
|
|
const looksLikeJson = dep.trimStart().startsWith('{');
|
|
|
|
if (isJson) {
|
|
if (looksLikeJson) {
|
|
const releases = PypiDatasource.getSimpleReleasesFromJson(
|
|
dep,
|
|
packageName,
|
|
);
|
|
// A parse failure is distinct from a package with genuinely no files, so don't report an empty release list as if it were a real (if empty) result, which would be cached as such.
|
|
if (!releases) {
|
|
return null;
|
|
}
|
|
dependency.releases = PypiDatasource.toReleases(releases);
|
|
return dependency;
|
|
}
|
|
|
|
// The registry mislabeled a non-JSON body as JSON, so fall through to the HTML parser instead of reporting the package as not found.
|
|
logger.debug(
|
|
{ packageName, hostUrl, contentType },
|
|
'Parsing Simple API response as HTML, as it is labeled as JSON but does not look like it',
|
|
);
|
|
}
|
|
|
|
const htmlReleases = PypiDatasource.getSimpleReleasesFromHtml(
|
|
dep,
|
|
packageName,
|
|
);
|
|
dependency.releases = PypiDatasource.toReleases(htmlReleases);
|
|
if (dependency.releases.length > 0 || !looksLikeJson) {
|
|
return dependency;
|
|
}
|
|
|
|
// The registry honored the `Accept` header but omitted or mislabeled the `content-type`, so a JSON body went through the HTML parser and found nothing. Parse it as JSON instead of silently reporting no releases.
|
|
logger.debug(
|
|
{ packageName, hostUrl, contentType },
|
|
'Retrying Simple API response as JSON, as it is not labeled as JSON but looks like it',
|
|
);
|
|
const jsonReleases = PypiDatasource.getSimpleReleasesFromJson(
|
|
dep,
|
|
packageName,
|
|
);
|
|
if (!jsonReleases) {
|
|
return null;
|
|
}
|
|
dependency.releases = PypiDatasource.toReleases(jsonReleases);
|
|
return dependency;
|
|
}
|
|
|
|
/**
|
|
* The files of a version are not ordered by upload time, so use the earliest one: that is when the version was first published, which is what `minimumReleaseAge` needs.
|
|
*/
|
|
private static toReleases(releases: Releases): Release[] {
|
|
const versions = Object.keys(releases);
|
|
return versions.map((version) => {
|
|
const versionReleases = coerceArray(releases[version]);
|
|
const isDeprecated = versionReleases.some(({ yanked }) => yanked);
|
|
const result: Release = { version };
|
|
const releaseTimestamp =
|
|
PypiDatasource.getEarliestTimestamp(versionReleases);
|
|
if (releaseTimestamp) {
|
|
result.releaseTimestamp = releaseTimestamp;
|
|
}
|
|
if (isDeprecated) {
|
|
result.isDeprecated = isDeprecated;
|
|
}
|
|
const pythonConstraints = versionReleases.map(
|
|
({ requires_python }) => requires_python,
|
|
);
|
|
// A file without `requires_python` can be installed on any Python version, so the version as a whole is unconstrained
|
|
const isUnconstrained = pythonConstraints.some(
|
|
(constraint) => !isNonEmptyString(constraint),
|
|
);
|
|
// There may be multiple releases with different requires_python, so we return all in an array.
|
|
// Report no constraints at all for an unconstrained version, instead of only those of its other files, which would drop the version under `constraintsFiltering=strict`.
|
|
result.constraints = {
|
|
python: isUnconstrained
|
|
? []
|
|
: deduplicateArray(pythonConstraints.filter(isNonEmptyString)),
|
|
};
|
|
return result;
|
|
});
|
|
}
|
|
}
|