@macula-io/martha 0.0.0 → 1.0.0

package/node_modules/cacache/lib/entry-index.js

@@ -0,0 +1,339 @@
+'use strict'
+
+const crypto = require('crypto')
+const {
+  appendFile,
+  mkdir,
+  readFile,
+  readdir,
+  rm,
+  writeFile,
+} = require('fs/promises')
+const { Minipass } = require('minipass')
+const path = require('path')
+const ssri = require('ssri')
+const { tmpName } = require('./util/tmp')
+const cacheDir = require('./util/cache-dir')
+
+const contentPath = require('./content/path')
+const hashToSegments = require('./util/hash-to-segments')
+const indexV = require('../package.json')['cache-version'].index
+const { moveFile } = require('@npmcli/fs')
+
+const lsStreamConcurrency = 5
+
+module.exports.NotFoundError = class NotFoundError extends Error {
+  constructor (cache, key) {
+    super(`No cache entry for ${key} found in ${cache}`)
+    this.code = 'ENOENT'
+    this.cache = cache
+    this.key = key
+  }
+}
+
+module.exports.compact = compact
+
+async function compact (cache, key, matchFn, opts = {}) {
+  const bucket = bucketPath(cache, key)
+  const entries = await bucketEntries(bucket)
+  const newEntries = []
+  // we loop backwards because the bottom-most result is the newest
+  // since we add new entries with appendFile
+  for (let i = entries.length - 1; i >= 0; --i) {
+    const entry = entries[i]
+    // a null integrity could mean either a delete was appended
+    // or the user has simply stored an index that does not map
+    // to any content. we determine if the user wants to keep the
+    // null integrity based on the validateEntry function passed in options.
+    // if the integrity is null and no validateEntry is provided, we break
+    // as we consider the null integrity to be a deletion of everything
+    // that came before it.
+    if (entry.integrity === null && !opts.validateEntry) {
+      break
+    }
+
+    // if this entry is valid, and it is either the first entry or
+    // the newEntries array doesn't already include an entry that
+    // matches this one based on the provided matchFn, then we add
+    // it to the beginning of our list
+    if ((!opts.validateEntry || opts.validateEntry(entry) === true) &&
+      (newEntries.length === 0 ||
+        !newEntries.find((oldEntry) => matchFn(oldEntry, entry)))) {
+      newEntries.unshift(entry)
+    }
+  }
+
+  const newIndex = '\n' + newEntries.map((entry) => {
+    const stringified = JSON.stringify(entry)
+    const hash = hashEntry(stringified)
+    return `${hash}\t${stringified}`
+  }).join('\n')
+
+  const setup = async () => {
+    const target = tmpName(cache, opts.tmpPrefix)
+    await cacheDir.mkdir(cache)
+    await mkdir(path.dirname(target), { recursive: true })
+    return {
+      target,
+      moved: false,
+    }
+  }
+
+  const teardown = async (tmp) => {
+    if (!tmp.moved) {
+      return rm(tmp.target, { recursive: true, force: true })
+    }
+  }
+
+  const write = async (tmp) => {
+    await writeFile(tmp.target, newIndex, { flag: 'wx' })
+    await mkdir(path.dirname(bucket), { recursive: true })
+    // we use @npmcli/move-file directly here because we
+    // want to overwrite the existing file
+    await moveFile(tmp.target, bucket)
+    tmp.moved = true
+  }
+
+  // write the file atomically
+  const tmp = await setup()
+  try {
+    await write(tmp)
+  } finally {
+    await teardown(tmp)
+  }
+
+  // we reverse the list we generated such that the newest
+  // entries come first in order to make looping through them easier
+  // the true passed to formatEntry tells it to keep null
+  // integrity values, if they made it this far it's because
+  // validateEntry returned true, and as such we should return it
+  return newEntries.reverse().map((entry) => formatEntry(cache, entry, true))
+}
+
+module.exports.insert = insert
+
+async function insert (cache, key, integrity, opts = {}) {
+  const { metadata, size, time } = opts
+  const bucket = bucketPath(cache, key)
+  const entry = {
+    key,
+    integrity: integrity && ssri.stringify(integrity),
+    time: time || Date.now(),
+    size,
+    metadata,
+  }
+  try {
+    await cacheDir.mkdir(cache)
+    await mkdir(path.dirname(bucket), { recursive: true })
+    const stringified = JSON.stringify(entry)
+    // NOTE - Cleverness ahoy!
+    //
+    // This works because it's tremendously unlikely for an entry to corrupt
+    // another while still preserving the string length of the JSON in
+    // question. So, we just slap the length in there and verify it on read.
+    //
+    // Thanks to @isaacs for the whiteboarding session that ended up with
+    // this.
+    await appendFile(bucket, `\n${hashEntry(stringified)}\t${stringified}`)
+  } catch (err) {
+    if (err.code === 'ENOENT') {
+      return undefined
+    }
+
+    throw err
+  }
+  return formatEntry(cache, entry)
+}
+
+module.exports.find = find
+
+async function find (cache, key) {
+  const bucket = bucketPath(cache, key)
+  try {
+    const entries = await bucketEntries(bucket)
+    return entries.reduce((latest, next) => {
+      if (next && next.key === key) {
+        return formatEntry(cache, next)
+      } else {
+        return latest
+      }
+    }, null)
+  } catch (err) {
+    if (err.code === 'ENOENT') {
+      return null
+    } else {
+      throw err
+    }
+  }
+}
+
+module.exports.delete = del
+
+function del (cache, key, opts = {}) {
+  if (!opts.removeFully) {
+    return insert(cache, key, null, opts)
+  }
+
+  const bucket = bucketPath(cache, key)
+  return rm(bucket, { recursive: true, force: true })
+}
+
+module.exports.lsStream = lsStream
+
+function lsStream (cache) {
+  const indexDir = bucketDir(cache)
+  const stream = new Minipass({ objectMode: true })
+
+  // Set all this up to run on the stream and then just return the stream
+  Promise.resolve().then(async () => {
+    const { default: pMap } = await import('p-map')
+    const buckets = await readdirOrEmpty(indexDir)
+    await pMap(buckets, async (bucket) => {
+      const bucketPath = path.join(indexDir, bucket)
+      const subbuckets = await readdirOrEmpty(bucketPath)
+      await pMap(subbuckets, async (subbucket) => {
+        const subbucketPath = path.join(bucketPath, subbucket)
+
+        // "/cachename/<bucket 0xFF>/<bucket 0xFF>./*"
+        const subbucketEntries = await readdirOrEmpty(subbucketPath)
+        await pMap(subbucketEntries, async (entry) => {
+          const entryPath = path.join(subbucketPath, entry)
+          try {
+            const entries = await bucketEntries(entryPath)
+            // using a Map here prevents duplicate keys from showing up
+            // twice, I guess?
+            const reduced = entries.reduce((acc, entry) => {
+              acc.set(entry.key, entry)
+              return acc
+            }, new Map())
+            // reduced is a map of key => entry
+            for (const entry of reduced.values()) {
+              const formatted = formatEntry(cache, entry)
+              if (formatted) {
+                stream.write(formatted)
+              }
+            }
+          } catch (err) {
+            if (err.code === 'ENOENT') {
+              return undefined
+            }
+            throw err
+          }
+        },
+        { concurrency: lsStreamConcurrency })
+      },
+      { concurrency: lsStreamConcurrency })
+    },
+    { concurrency: lsStreamConcurrency })
+    stream.end()
+    return stream
+  }).catch(err => stream.emit('error', err))
+
+  return stream
+}
+
+module.exports.ls = ls
+
+async function ls (cache) {
+  const entries = await lsStream(cache).collect()
+  return entries.reduce((acc, xs) => {
+    acc[xs.key] = xs
+    return acc
+  }, {})
+}
+
+module.exports.bucketEntries = bucketEntries
+
+async function bucketEntries (bucket, filter) {
+  const data = await readFile(bucket, 'utf8')
+  return _bucketEntries(data, filter)
+}
+
+function _bucketEntries (data) {
+  const entries = []
+  data.split('\n').forEach((entry) => {
+    if (!entry) {
+      return
+    }
+
+    const pieces = entry.split('\t')
+    if (!pieces[1] || hashEntry(pieces[1]) !== pieces[0]) {
+      // Hash is no good! Corruption or malice? Doesn't matter!
+      // EJECT EJECT
+      return
+    }
+    let obj
+    try {
+      obj = JSON.parse(pieces[1])
+    } catch (_) {
+      // eslint-ignore-next-line no-empty-block
+    }
+    // coverage disabled here, no need to test with an entry that parses to something falsey
+    // istanbul ignore else
+    if (obj) {
+      entries.push(obj)
+    }
+  })
+  return entries
+}
+
+module.exports.bucketDir = bucketDir
+
+function bucketDir (cache) {
+  return path.join(cache, `index-v${indexV}`)
+}
+
+module.exports.bucketPath = bucketPath
+
+function bucketPath (cache, key) {
+  const hashed = hashKey(key)
+  return path.join.apply(
+    path,
+    [bucketDir(cache)].concat(hashToSegments(hashed))
+  )
+}
+
+module.exports.hashKey = hashKey
+
+function hashKey (key) {
+  return hash(key, 'sha256')
+}
+
+module.exports.hashEntry = hashEntry
+
+function hashEntry (str) {
+  return hash(str, 'sha1')
+}
+
+function hash (str, digest) {
+  return crypto
+    .createHash(digest)
+    .update(str)
+    .digest('hex')
+}
+
+function formatEntry (cache, entry, keepAll) {
+  // Treat null digests as deletions. They'll shadow any previous entries.
+  if (!entry.integrity && !keepAll) {
+    return null
+  }
+
+  return {
+    key: entry.key,
+    integrity: entry.integrity,
+    path: entry.integrity ? contentPath(cache, entry.integrity) : undefined,
+    size: entry.size,
+    time: entry.time,
+    metadata: entry.metadata,
+  }
+}
+
+function readdirOrEmpty (dir) {
+  return readdir(dir).catch((err) => {
+    if (err.code === 'ENOENT' || err.code === 'ENOTDIR') {
+      return []
+    }
+
+    throw err
+  })
+}

package/node_modules/cacache/lib/get.js

@@ -0,0 +1,170 @@
+'use strict'
+
+const Collect = require('minipass-collect')
+const { Minipass } = require('minipass')
+const Pipeline = require('minipass-pipeline')
+
+const index = require('./entry-index')
+const memo = require('./memoization')
+const read = require('./content/read')
+
+async function getData (cache, key, opts = {}) {
+  const { integrity, memoize, size } = opts
+  const memoized = memo.get(cache, key, opts)
+  if (memoized && memoize !== false) {
+    return {
+      metadata: memoized.entry.metadata,
+      data: memoized.data,
+      integrity: memoized.entry.integrity,
+      size: memoized.entry.size,
+    }
+  }
+
+  const entry = await index.find(cache, key, opts)
+  if (!entry) {
+    throw new index.NotFoundError(cache, key)
+  }
+  const data = await read(cache, entry.integrity, { integrity, size })
+  if (memoize) {
+    memo.put(cache, entry, data, opts)
+  }
+
+  return {
+    data,
+    metadata: entry.metadata,
+    size: entry.size,
+    integrity: entry.integrity,
+  }
+}
+module.exports = getData
+
+async function getDataByDigest (cache, key, opts = {}) {
+  const { integrity, memoize, size } = opts
+  const memoized = memo.get.byDigest(cache, key, opts)
+  if (memoized && memoize !== false) {
+    return memoized
+  }
+
+  const res = await read(cache, key, { integrity, size })
+  if (memoize) {
+    memo.put.byDigest(cache, key, res, opts)
+  }
+  return res
+}
+module.exports.byDigest = getDataByDigest
+
+const getMemoizedStream = (memoized) => {
+  const stream = new Minipass()
+  stream.on('newListener', function (ev, cb) {
+    ev === 'metadata' && cb(memoized.entry.metadata)
+    ev === 'integrity' && cb(memoized.entry.integrity)
+    ev === 'size' && cb(memoized.entry.size)
+  })
+  stream.end(memoized.data)
+  return stream
+}
+
+function getStream (cache, key, opts = {}) {
+  const { memoize, size } = opts
+  const memoized = memo.get(cache, key, opts)
+  if (memoized && memoize !== false) {
+    return getMemoizedStream(memoized)
+  }
+
+  const stream = new Pipeline()
+  // Set all this up to run on the stream and then just return the stream
+  Promise.resolve().then(async () => {
+    const entry = await index.find(cache, key)
+    if (!entry) {
+      throw new index.NotFoundError(cache, key)
+    }
+
+    stream.emit('metadata', entry.metadata)
+    stream.emit('integrity', entry.integrity)
+    stream.emit('size', entry.size)
+    stream.on('newListener', function (ev, cb) {
+      ev === 'metadata' && cb(entry.metadata)
+      ev === 'integrity' && cb(entry.integrity)
+      ev === 'size' && cb(entry.size)
+    })
+
+    const src = read.readStream(
+      cache,
+      entry.integrity,
+      { ...opts, size: typeof size !== 'number' ? entry.size : size }
+    )
+
+    if (memoize) {
+      const memoStream = new Collect.PassThrough()
+      memoStream.on('collect', data => memo.put(cache, entry, data, opts))
+      stream.unshift(memoStream)
+    }
+    stream.unshift(src)
+    return stream
+  }).catch((err) => stream.emit('error', err))
+
+  return stream
+}
+
+module.exports.stream = getStream
+
+function getStreamDigest (cache, integrity, opts = {}) {
+  const { memoize } = opts
+  const memoized = memo.get.byDigest(cache, integrity, opts)
+  if (memoized && memoize !== false) {
+    const stream = new Minipass()
+    stream.end(memoized)
+    return stream
+  } else {
+    const stream = read.readStream(cache, integrity, opts)
+    if (!memoize) {
+      return stream
+    }
+
+    const memoStream = new Collect.PassThrough()
+    memoStream.on('collect', data => memo.put.byDigest(
+      cache,
+      integrity,
+      data,
+      opts
+    ))
+    return new Pipeline(stream, memoStream)
+  }
+}
+
+module.exports.stream.byDigest = getStreamDigest
+
+function info (cache, key, opts = {}) {
+  const { memoize } = opts
+  const memoized = memo.get(cache, key, opts)
+  if (memoized && memoize !== false) {
+    return Promise.resolve(memoized.entry)
+  } else {
+    return index.find(cache, key)
+  }
+}
+module.exports.info = info
+
+async function copy (cache, key, dest, opts = {}) {
+  const entry = await index.find(cache, key, opts)
+  if (!entry) {
+    throw new index.NotFoundError(cache, key)
+  }
+  await read.copy(cache, entry.integrity, dest, opts)
+  return {
+    metadata: entry.metadata,
+    size: entry.size,
+    integrity: entry.integrity,
+  }
+}
+
+module.exports.copy = copy
+
+async function copyByDigest (cache, key, dest, opts = {}) {
+  await read.copy(cache, key, dest, opts)
+  return key
+}
+
+module.exports.copy.byDigest = copyByDigest
+
+module.exports.hasContent = read.hasContent

package/node_modules/cacache/lib/index.js

@@ -0,0 +1,42 @@
+'use strict'
+
+const get = require('./get.js')
+const put = require('./put.js')
+const rm = require('./rm.js')
+const verify = require('./verify.js')
+const { clearMemoized } = require('./memoization.js')
+const tmp = require('./util/tmp.js')
+const index = require('./entry-index.js')
+
+module.exports.index = {}
+module.exports.index.compact = index.compact
+module.exports.index.insert = index.insert
+
+module.exports.ls = index.ls
+module.exports.ls.stream = index.lsStream
+
+module.exports.get = get
+module.exports.get.byDigest = get.byDigest
+module.exports.get.stream = get.stream
+module.exports.get.stream.byDigest = get.stream.byDigest
+module.exports.get.copy = get.copy
+module.exports.get.copy.byDigest = get.copy.byDigest
+module.exports.get.info = get.info
+module.exports.get.hasContent = get.hasContent
+
+module.exports.put = put
+module.exports.put.stream = put.stream
+
+module.exports.rm = rm.entry
+module.exports.rm.all = rm.all
+module.exports.rm.entry = module.exports.rm
+module.exports.rm.content = rm.content
+
+module.exports.clearMemoized = clearMemoized
+
+module.exports.tmp = {}
+module.exports.tmp.mkdir = tmp.mkdir
+module.exports.tmp.withTmp = tmp.withTmp
+
+module.exports.verify = verify
+module.exports.verify.lastRun = verify.lastRun

package/node_modules/cacache/lib/memoization.js

@@ -0,0 +1,72 @@
+'use strict'
+
+const { LRUCache } = require('lru-cache')
+
+const MEMOIZED = new LRUCache({
+  max: 500,
+  maxSize: 50 * 1024 * 1024, // 50MB
+  ttl: 3 * 60 * 1000, // 3 minutes
+  sizeCalculation: (entry, key) => key.startsWith('key:') ? entry.data.length : entry.length,
+})
+
+module.exports.clearMemoized = clearMemoized
+
+function clearMemoized () {
+  const old = {}
+  MEMOIZED.forEach((v, k) => {
+    old[k] = v
+  })
+  MEMOIZED.clear()
+  return old
+}
+
+module.exports.put = put
+
+function put (cache, entry, data, opts) {
+  pickMem(opts).set(`key:${cache}:${entry.key}`, { entry, data })
+  putDigest(cache, entry.integrity, data, opts)
+}
+
+module.exports.put.byDigest = putDigest
+
+function putDigest (cache, integrity, data, opts) {
+  pickMem(opts).set(`digest:${cache}:${integrity}`, data)
+}
+
+module.exports.get = get
+
+function get (cache, key, opts) {
+  return pickMem(opts).get(`key:${cache}:${key}`)
+}
+
+module.exports.get.byDigest = getDigest
+
+function getDigest (cache, integrity, opts) {
+  return pickMem(opts).get(`digest:${cache}:${integrity}`)
+}
+
+class ObjProxy {
+  constructor (obj) {
+    this.obj = obj
+  }
+
+  get (key) {
+    return this.obj[key]
+  }
+
+  set (key, val) {
+    this.obj[key] = val
+  }
+}
+
+function pickMem (opts) {
+  if (!opts || !opts.memoize) {
+    return MEMOIZED
+  } else if (opts.memoize.get && opts.memoize.set) {
+    return opts.memoize
+  } else if (typeof opts.memoize === 'object') {
+    return new ObjProxy(opts.memoize)
+  } else {
+    return MEMOIZED
+  }
+}

package/node_modules/cacache/lib/put.js

@@ -0,0 +1,80 @@
+'use strict'
+
+const index = require('./entry-index')
+const memo = require('./memoization')
+const write = require('./content/write')
+const Flush = require('minipass-flush')
+const { PassThrough } = require('minipass-collect')
+const Pipeline = require('minipass-pipeline')
+
+const putOpts = (opts) => ({
+  algorithms: ['sha512'],
+  ...opts,
+})
+
+module.exports = putData
+
+async function putData (cache, key, data, opts = {}) {
+  const { memoize } = opts
+  opts = putOpts(opts)
+  const res = await write(cache, data, opts)
+  const entry = await index.insert(cache, key, res.integrity, { ...opts, size: res.size })
+  if (memoize) {
+    memo.put(cache, entry, data, opts)
+  }
+
+  return res.integrity
+}
+
+module.exports.stream = putStream
+
+function putStream (cache, key, opts = {}) {
+  const { memoize } = opts
+  opts = putOpts(opts)
+  let integrity
+  let size
+  let error
+
+  let memoData
+  const pipeline = new Pipeline()
+  // first item in the pipeline is the memoizer, because we need
+  // that to end first and get the collected data.
+  if (memoize) {
+    const memoizer = new PassThrough().on('collect', data => {
+      memoData = data
+    })
+    pipeline.push(memoizer)
+  }
+
+  // contentStream is a write-only, not a passthrough
+  // no data comes out of it.
+  const contentStream = write.stream(cache, opts)
+    .on('integrity', (int) => {
+      integrity = int
+    })
+    .on('size', (s) => {
+      size = s
+    })
+    .on('error', (err) => {
+      error = err
+    })
+
+  pipeline.push(contentStream)
+
+  // last but not least, we write the index and emit hash and size,
+  // and memoize if we're doing that
+  pipeline.push(new Flush({
+    async flush () {
+      if (!error) {
+        const entry = await index.insert(cache, key, integrity, { ...opts, size })
+        if (memoize && memoData) {
+          memo.put(cache, entry, memoData, opts)
+        }
+        pipeline.emit('integrity', integrity)
+        pipeline.emit('size', size)
+      }
+    },
+  }))
+
+  return pipeline
+}