@naturalcycles/nodejs-lib 12.55.1 → 12.56.0
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/dist/index.d.ts +2 -2
- package/dist/index.js +3 -3
- package/dist/stream/ndjson/ndJsonFileRead.js +1 -1
- package/dist/stream/transform/{beta/transformMap2.d.ts → legacy/transformMap.d.ts} +1 -1
- package/dist/stream/transform/legacy/transformMap.js +94 -0
- package/dist/stream/transform/transformMap.js +48 -44
- package/dist/stream/writable/writableForEach.d.ts +5 -1
- package/dist/stream/writable/writableForEach.js +8 -1
- package/package.json +1 -1
- package/src/index.ts +1 -2
- package/src/stream/ndjson/ndJsonFileRead.ts +2 -2
- package/src/stream/transform/legacy/transformMap.ts +133 -0
- package/src/stream/transform/transformMap.ts +67 -66
- package/src/stream/writable/writableForEach.ts +12 -2
- package/dist/stream/transform/beta/transformMap2.js +0 -98
- package/src/stream/transform/beta/transformMap2.ts +0 -134
package/dist/index.d.ts
CHANGED
|
@@ -47,7 +47,7 @@ import { transformToString } from './stream/transform/transformToString';
|
|
|
47
47
|
import { BaseWorkerClass, WorkerClassInterface } from './stream/transform/worker/baseWorkerClass';
|
|
48
48
|
import { transformMultiThreaded, TransformMultiThreadedOptions } from './stream/transform/worker/transformMultiThreaded';
|
|
49
49
|
import { WorkerInput, WorkerOutput } from './stream/transform/worker/transformMultiThreaded.model';
|
|
50
|
-
|
|
50
|
+
export * from './stream/writable/writableForEach';
|
|
51
51
|
import { writableFork } from './stream/writable/writableFork';
|
|
52
52
|
import { writablePushToArray } from './stream/writable/writablePushToArray';
|
|
53
53
|
import { writableVoid } from './stream/writable/writableVoid';
|
|
@@ -67,4 +67,4 @@ import { convert, getValidationResult, isValid, JoiValidationResult, undefinedIf
|
|
|
67
67
|
import { sanitizeHTML, SanitizeHTMLOptions } from './validation/sanitize.util';
|
|
68
68
|
import { runScript, RunScriptOptions } from './script';
|
|
69
69
|
export type { RunScriptOptions, JoiValidationErrorData, JoiValidationResult, ValidationErrorItem, ExtendedJoi, SchemaTyped, AnySchema, AnySchemaTyped, ArraySchemaTyped, BooleanSchemaTyped, NumberSchemaTyped, ObjectSchemaTyped, StringSchemaTyped, IDebug, IDebugger, SlackServiceCfg, SlackMessage, SlackMessageProps, SlackApiBody, SlackMessagePrefixHook, ReadableTyped, WritableTyped, TransformTyped, PipelineFromNDJsonFileOptions, PipelineToNDJsonFileOptions, TransformJsonParseOptions, TransformToNDJsonOptions, TransformMapOptions, TransformMapSyncOptions, NDJSONStreamForEachOptions, TransformOptions, TransformLogProgressOptions, TransformMultiThreadedOptions, WorkerClassInterface, WorkerInput, WorkerOutput, TableDiffOptions, InspectAnyOptions, Got, GetGotOptions, AfterResponseHook, BeforeErrorHook, BeforeRequestHook, AjvValidationOptions, AjvSchemaCfg, AjvValidationErrorData, SanitizeHTMLOptions, };
|
|
70
|
-
export { JoiValidationError, validate, getValidationResult, isValid, undefinedIfInvalid, convert, Joi, booleanSchema, booleanDefaultToFalseSchema, stringSchema, numberSchema, integerSchema, percentageSchema, dateStringSchema, arraySchema, binarySchema, objectSchema, oneOfSchema, anySchema, anyObjectSchema, baseDBEntitySchema, savedDBEntitySchema, idSchema, unixTimestampSchema, verSchema, emailSchema, SEM_VER_PATTERN, semVerSchema, userAgentSchema, utcOffsetSchema, ipAddressSchema, slugSchema, urlSchema, processSharedUtil, zipBuffer, gzipBuffer, unzipBuffer, gunzipBuffer, zipString, gzipString, unzipToString, gunzipToString, requireEnvKeys, requireFileToExist, LRUMemoCache, stringId, stringIdAsync, stringIdUnsafe, ALPHABET_NUMBER, ALPHABET_LOWERCASE, ALPHABET_UPPERCASE, ALPHABET_ALPHANUMERIC_LOWERCASE, ALPHABET_ALPHANUMERIC_UPPERCASE, ALPHABET_ALPHANUMERIC, md5, hash, hashAsBuffer, md5AsBuffer, stringToBase64, base64ToString, bufferToBase64, base64ToBuffer, Debug, getSecretMap, setSecretMap, loadSecretsFromEnv, loadSecretsFromJsonFile, removeSecretsFromEnv, secret, secretOptional, memoryUsage, memoryUsageFull, SlackService, slackDefaultMessagePrefixHook, readableCreate, readableFrom, readableFromArray, readableToArray, readableForEach, readableForEachSync, readableMap, readableMapToArray, _pipeline, transformBuffer, ndjsonMap, ndJsonFileRead, ndJsonFileWrite, ndjsonStreamForEach, pipelineFromNDJsonFile, pipelineToNDJsonFile, NDJsonStats, streamToNDJsonFile, transformJsonParse, bufferReviver, transformToNDJson, transformFilter, transformFilterSync, transformMap, transformMapSync, transformMapSimple, transformNoOp,
|
|
70
|
+
export { JoiValidationError, validate, getValidationResult, isValid, undefinedIfInvalid, convert, Joi, booleanSchema, booleanDefaultToFalseSchema, stringSchema, numberSchema, integerSchema, percentageSchema, dateStringSchema, arraySchema, binarySchema, objectSchema, oneOfSchema, anySchema, anyObjectSchema, baseDBEntitySchema, savedDBEntitySchema, idSchema, unixTimestampSchema, verSchema, emailSchema, SEM_VER_PATTERN, semVerSchema, userAgentSchema, utcOffsetSchema, ipAddressSchema, slugSchema, urlSchema, processSharedUtil, zipBuffer, gzipBuffer, unzipBuffer, gunzipBuffer, zipString, gzipString, unzipToString, gunzipToString, requireEnvKeys, requireFileToExist, LRUMemoCache, stringId, stringIdAsync, stringIdUnsafe, ALPHABET_NUMBER, ALPHABET_LOWERCASE, ALPHABET_UPPERCASE, ALPHABET_ALPHANUMERIC_LOWERCASE, ALPHABET_ALPHANUMERIC_UPPERCASE, ALPHABET_ALPHANUMERIC, md5, hash, hashAsBuffer, md5AsBuffer, stringToBase64, base64ToString, bufferToBase64, base64ToBuffer, Debug, getSecretMap, setSecretMap, loadSecretsFromEnv, loadSecretsFromJsonFile, removeSecretsFromEnv, secret, secretOptional, memoryUsage, memoryUsageFull, SlackService, slackDefaultMessagePrefixHook, readableCreate, readableFrom, readableFromArray, readableToArray, readableForEach, readableForEachSync, readableMap, readableMapToArray, _pipeline, transformBuffer, ndjsonMap, ndJsonFileRead, ndJsonFileWrite, ndjsonStreamForEach, pipelineFromNDJsonFile, pipelineToNDJsonFile, NDJsonStats, streamToNDJsonFile, transformJsonParse, bufferReviver, transformToNDJson, transformFilter, transformFilterSync, transformMap, transformMapSync, transformMapSimple, transformNoOp, writablePushToArray, transformSplit, transformToString, transformToArray, transformTap, transformLogProgress, transformLimit, writableVoid, writableFork, transformMultiThreaded, BaseWorkerClass, tableDiff, inspectAny, inspectAnyStringifyFn, getGot, HTTPError, TimeoutError, _chunkBuffer, Ajv, getAjv, AjvSchema, AjvValidationError, readJsonSchemas, readAjvSchemas, hasColors, sanitizeHTML, runScript, };
|
package/dist/index.js
CHANGED
|
@@ -2,7 +2,8 @@
|
|
|
2
2
|
Object.defineProperty(exports, "__esModule", { value: true });
|
|
3
3
|
exports.ALPHABET_LOWERCASE = exports.ALPHABET_NUMBER = exports.stringIdUnsafe = exports.stringIdAsync = exports.stringId = exports.LRUMemoCache = exports.requireFileToExist = exports.requireEnvKeys = exports.gunzipToString = exports.unzipToString = exports.gzipString = exports.zipString = exports.gunzipBuffer = exports.unzipBuffer = exports.gzipBuffer = exports.zipBuffer = exports.processSharedUtil = exports.urlSchema = exports.slugSchema = exports.ipAddressSchema = exports.utcOffsetSchema = exports.userAgentSchema = exports.semVerSchema = exports.SEM_VER_PATTERN = exports.emailSchema = exports.verSchema = exports.unixTimestampSchema = exports.idSchema = exports.savedDBEntitySchema = exports.baseDBEntitySchema = exports.anyObjectSchema = exports.anySchema = exports.oneOfSchema = exports.objectSchema = exports.binarySchema = exports.arraySchema = exports.dateStringSchema = exports.percentageSchema = exports.integerSchema = exports.numberSchema = exports.stringSchema = exports.booleanDefaultToFalseSchema = exports.booleanSchema = exports.Joi = exports.convert = exports.undefinedIfInvalid = exports.isValid = exports.getValidationResult = exports.validate = exports.JoiValidationError = void 0;
|
|
4
4
|
exports.transformMapSimple = exports.transformMapSync = exports.transformMap = exports.transformFilterSync = exports.transformFilter = exports.transformToNDJson = exports.bufferReviver = exports.transformJsonParse = exports.streamToNDJsonFile = exports.NDJsonStats = exports.pipelineToNDJsonFile = exports.pipelineFromNDJsonFile = exports.ndjsonStreamForEach = exports.ndJsonFileWrite = exports.ndJsonFileRead = exports.ndjsonMap = exports.transformBuffer = exports._pipeline = exports.readableMapToArray = exports.readableMap = exports.readableForEachSync = exports.readableForEach = exports.readableToArray = exports.readableFromArray = exports.readableFrom = exports.readableCreate = exports.slackDefaultMessagePrefixHook = exports.SlackService = exports.memoryUsageFull = exports.memoryUsage = exports.secretOptional = exports.secret = exports.removeSecretsFromEnv = exports.loadSecretsFromJsonFile = exports.loadSecretsFromEnv = exports.setSecretMap = exports.getSecretMap = exports.Debug = exports.base64ToBuffer = exports.bufferToBase64 = exports.base64ToString = exports.stringToBase64 = exports.md5AsBuffer = exports.hashAsBuffer = exports.hash = exports.md5 = exports.ALPHABET_ALPHANUMERIC = exports.ALPHABET_ALPHANUMERIC_UPPERCASE = exports.ALPHABET_ALPHANUMERIC_LOWERCASE = exports.ALPHABET_UPPERCASE = void 0;
|
|
5
|
-
exports.runScript = exports.sanitizeHTML = exports.hasColors = exports.readAjvSchemas = exports.readJsonSchemas = exports.AjvValidationError = exports.AjvSchema = exports.getAjv = exports.Ajv = exports._chunkBuffer = exports.TimeoutError = exports.HTTPError = exports.getGot = exports.inspectAnyStringifyFn = exports.inspectAny = exports.tableDiff = exports.BaseWorkerClass = exports.transformMultiThreaded = exports.writableFork = exports.writableVoid = exports.transformLimit = exports.transformLogProgress = exports.transformTap = exports.transformToArray = exports.transformToString = exports.transformSplit = exports.writablePushToArray = exports.
|
|
5
|
+
exports.runScript = exports.sanitizeHTML = exports.hasColors = exports.readAjvSchemas = exports.readJsonSchemas = exports.AjvValidationError = exports.AjvSchema = exports.getAjv = exports.Ajv = exports._chunkBuffer = exports.TimeoutError = exports.HTTPError = exports.getGot = exports.inspectAnyStringifyFn = exports.inspectAny = exports.tableDiff = exports.BaseWorkerClass = exports.transformMultiThreaded = exports.writableFork = exports.writableVoid = exports.transformLimit = exports.transformLogProgress = exports.transformTap = exports.transformToArray = exports.transformToString = exports.transformSplit = exports.writablePushToArray = exports.transformNoOp = void 0;
|
|
6
|
+
const tslib_1 = require("tslib");
|
|
6
7
|
const ajv_1 = require("ajv");
|
|
7
8
|
exports.Ajv = ajv_1.default;
|
|
8
9
|
const got_1 = require("got");
|
|
@@ -118,8 +119,7 @@ const baseWorkerClass_1 = require("./stream/transform/worker/baseWorkerClass");
|
|
|
118
119
|
Object.defineProperty(exports, "BaseWorkerClass", { enumerable: true, get: function () { return baseWorkerClass_1.BaseWorkerClass; } });
|
|
119
120
|
const transformMultiThreaded_1 = require("./stream/transform/worker/transformMultiThreaded");
|
|
120
121
|
Object.defineProperty(exports, "transformMultiThreaded", { enumerable: true, get: function () { return transformMultiThreaded_1.transformMultiThreaded; } });
|
|
121
|
-
|
|
122
|
-
Object.defineProperty(exports, "writableForEach", { enumerable: true, get: function () { return writableForEach_1.writableForEach; } });
|
|
122
|
+
(0, tslib_1.__exportStar)(require("./stream/writable/writableForEach"), exports);
|
|
123
123
|
const writableFork_1 = require("./stream/writable/writableFork");
|
|
124
124
|
Object.defineProperty(exports, "writableFork", { enumerable: true, get: function () { return writableFork_1.writableFork; } });
|
|
125
125
|
const writablePushToArray_1 = require("./stream/writable/writablePushToArray");
|
|
@@ -8,7 +8,7 @@ const pipelineFromNDJsonFile_1 = require("./pipelineFromNDJsonFile");
|
|
|
8
8
|
*/
|
|
9
9
|
async function ndJsonFileRead(opt) {
|
|
10
10
|
const res = [];
|
|
11
|
-
await (0, pipelineFromNDJsonFile_1.pipelineFromNDJsonFile)([(0, __1.
|
|
11
|
+
await (0, pipelineFromNDJsonFile_1.pipelineFromNDJsonFile)([(0, __1.writablePushToArray)(res)], opt);
|
|
12
12
|
return res;
|
|
13
13
|
}
|
|
14
14
|
exports.ndJsonFileRead = ndJsonFileRead;
|
|
@@ -14,4 +14,4 @@ export declare function notNullishPredicate(item: any): boolean;
|
|
|
14
14
|
*
|
|
15
15
|
* If an Array is returned by `mapper` - it will be flattened and multiple results will be emitted from it. Tested by Array.isArray().
|
|
16
16
|
*/
|
|
17
|
-
export declare function
|
|
17
|
+
export declare function transformMapLegacy<IN = any, OUT = IN>(mapper: AsyncMapper<IN, OUT>, opt?: TransformMapOptions<IN, OUT>): TransformTyped<IN, OUT>;
|
|
@@ -0,0 +1,94 @@
|
|
|
1
|
+
"use strict";
|
|
2
|
+
Object.defineProperty(exports, "__esModule", { value: true });
|
|
3
|
+
exports.transformMapLegacy = exports.notNullishPredicate = void 0;
|
|
4
|
+
const js_lib_1 = require("@naturalcycles/js-lib");
|
|
5
|
+
const through2Concurrent = require("through2-concurrent");
|
|
6
|
+
const colors_1 = require("../../../colors");
|
|
7
|
+
function notNullishPredicate(item) {
|
|
8
|
+
return item !== undefined && item !== null;
|
|
9
|
+
}
|
|
10
|
+
exports.notNullishPredicate = notNullishPredicate;
|
|
11
|
+
/**
|
|
12
|
+
* Like pMap, but for streams.
|
|
13
|
+
* Inspired by `through2`.
|
|
14
|
+
* Main feature is concurrency control (implemented via `through2-concurrent`) and convenient options.
|
|
15
|
+
* Using this allows native stream .pipe() to work and use backpressure.
|
|
16
|
+
*
|
|
17
|
+
* Only works in objectMode (due to through2Concurrent).
|
|
18
|
+
*
|
|
19
|
+
* Concurrency defaults to 16.
|
|
20
|
+
*
|
|
21
|
+
* If an Array is returned by `mapper` - it will be flattened and multiple results will be emitted from it. Tested by Array.isArray().
|
|
22
|
+
*/
|
|
23
|
+
function transformMapLegacy(mapper, opt = {}) {
|
|
24
|
+
const { concurrency = 16, predicate = notNullishPredicate, errorMode = js_lib_1.ErrorMode.THROW_IMMEDIATELY, flattenArrayOutput, onError, beforeFinal, metric = 'stream', logger = console, } = opt;
|
|
25
|
+
let index = -1;
|
|
26
|
+
let isRejected = false;
|
|
27
|
+
let errors = 0;
|
|
28
|
+
const collectedErrors = []; // only used if errorMode == THROW_AGGREGATED
|
|
29
|
+
return through2Concurrent.obj({
|
|
30
|
+
maxConcurrency: concurrency,
|
|
31
|
+
// autoDestroy: true,
|
|
32
|
+
async final(cb) {
|
|
33
|
+
// console.log('transformMap final')
|
|
34
|
+
logErrorStats(logger, true);
|
|
35
|
+
await beforeFinal?.(); // call beforeFinal if defined
|
|
36
|
+
if (collectedErrors.length) {
|
|
37
|
+
// emit Aggregated error
|
|
38
|
+
cb(new js_lib_1.AggregatedError(collectedErrors));
|
|
39
|
+
}
|
|
40
|
+
else {
|
|
41
|
+
// emit no error
|
|
42
|
+
cb();
|
|
43
|
+
}
|
|
44
|
+
},
|
|
45
|
+
}, async function transformMapFn(chunk, _encoding, cb) {
|
|
46
|
+
index++;
|
|
47
|
+
// console.log({chunk, _encoding})
|
|
48
|
+
// Stop processing if THROW_IMMEDIATELY mode is used
|
|
49
|
+
if (isRejected && errorMode === js_lib_1.ErrorMode.THROW_IMMEDIATELY)
|
|
50
|
+
return cb();
|
|
51
|
+
try {
|
|
52
|
+
const currentIndex = index; // because we need to pass it to 2 functions - mapper and predicate. Refers to INPUT index (since it may return multiple outputs)
|
|
53
|
+
const res = await mapper(chunk, currentIndex);
|
|
54
|
+
const passedResults = await (0, js_lib_1.pFilter)(flattenArrayOutput && Array.isArray(res) ? res : [res], async (r) => await predicate(r, currentIndex));
|
|
55
|
+
if (passedResults.length === 0) {
|
|
56
|
+
cb(); // 0 results
|
|
57
|
+
}
|
|
58
|
+
else {
|
|
59
|
+
passedResults.forEach(r => {
|
|
60
|
+
this.push(r);
|
|
61
|
+
// cb(null, r)
|
|
62
|
+
});
|
|
63
|
+
cb(); // done processing
|
|
64
|
+
}
|
|
65
|
+
}
|
|
66
|
+
catch (err) {
|
|
67
|
+
logger.error(err);
|
|
68
|
+
errors++;
|
|
69
|
+
logErrorStats(logger);
|
|
70
|
+
if (onError) {
|
|
71
|
+
try {
|
|
72
|
+
onError(err, chunk);
|
|
73
|
+
}
|
|
74
|
+
catch { }
|
|
75
|
+
}
|
|
76
|
+
if (errorMode === js_lib_1.ErrorMode.THROW_IMMEDIATELY) {
|
|
77
|
+
isRejected = true;
|
|
78
|
+
// Emit error immediately
|
|
79
|
+
return cb(err);
|
|
80
|
+
}
|
|
81
|
+
if (errorMode === js_lib_1.ErrorMode.THROW_AGGREGATED) {
|
|
82
|
+
collectedErrors.push(err);
|
|
83
|
+
}
|
|
84
|
+
// Tell input stream that we're done processing, but emit nothing to output - not error nor result
|
|
85
|
+
cb();
|
|
86
|
+
}
|
|
87
|
+
});
|
|
88
|
+
function logErrorStats(logger, final = false) {
|
|
89
|
+
if (!errors)
|
|
90
|
+
return;
|
|
91
|
+
logger.log(`${metric} ${final ? 'final ' : ''}errors: ${(0, colors_1.yellow)(errors)}`);
|
|
92
|
+
}
|
|
93
|
+
}
|
|
94
|
+
exports.transformMapLegacy = transformMapLegacy;
|
|
@@ -1,8 +1,8 @@
|
|
|
1
1
|
"use strict";
|
|
2
2
|
Object.defineProperty(exports, "__esModule", { value: true });
|
|
3
3
|
exports.transformMap = exports.notNullishPredicate = void 0;
|
|
4
|
+
const stream_1 = require("stream");
|
|
4
5
|
const js_lib_1 = require("@naturalcycles/js-lib");
|
|
5
|
-
const through2Concurrent = require("through2-concurrent");
|
|
6
6
|
const colors_1 = require("../../colors");
|
|
7
7
|
function notNullishPredicate(item) {
|
|
8
8
|
return item !== undefined && item !== null;
|
|
@@ -26,11 +26,17 @@ function transformMap(mapper, opt = {}) {
|
|
|
26
26
|
let isRejected = false;
|
|
27
27
|
let errors = 0;
|
|
28
28
|
const collectedErrors = []; // only used if errorMode == THROW_AGGREGATED
|
|
29
|
-
|
|
30
|
-
|
|
31
|
-
|
|
29
|
+
const q = new js_lib_1.PQueue({
|
|
30
|
+
concurrency,
|
|
31
|
+
resolveOn: 'start',
|
|
32
|
+
// debug: true,
|
|
33
|
+
});
|
|
34
|
+
return new stream_1.Transform({
|
|
35
|
+
objectMode: true,
|
|
32
36
|
async final(cb) {
|
|
33
|
-
// console.log('transformMap final')
|
|
37
|
+
// console.log('transformMap final', {index}, q.inFlight, q.queueSize)
|
|
38
|
+
// wait for the current inFlight jobs to complete and push their results
|
|
39
|
+
await q.onIdle();
|
|
34
40
|
logErrorStats(logger, true);
|
|
35
41
|
await beforeFinal?.(); // call beforeFinal if defined
|
|
36
42
|
if (collectedErrors.length) {
|
|
@@ -42,48 +48,46 @@ function transformMap(mapper, opt = {}) {
|
|
|
42
48
|
cb();
|
|
43
49
|
}
|
|
44
50
|
},
|
|
45
|
-
|
|
46
|
-
|
|
47
|
-
|
|
48
|
-
|
|
49
|
-
|
|
50
|
-
|
|
51
|
-
|
|
52
|
-
|
|
53
|
-
|
|
54
|
-
const passedResults = await (0, js_lib_1.pFilter)(flattenArrayOutput && Array.isArray(res) ? res : [res], async (r) => await predicate(r, currentIndex));
|
|
55
|
-
if (passedResults.length === 0) {
|
|
56
|
-
cb(); // 0 results
|
|
57
|
-
}
|
|
58
|
-
else {
|
|
59
|
-
passedResults.forEach(r => {
|
|
60
|
-
this.push(r);
|
|
61
|
-
// cb(null, r)
|
|
62
|
-
});
|
|
63
|
-
cb(); // done processing
|
|
64
|
-
}
|
|
65
|
-
}
|
|
66
|
-
catch (err) {
|
|
67
|
-
logger.error(err);
|
|
68
|
-
errors++;
|
|
69
|
-
logErrorStats(logger);
|
|
70
|
-
if (onError) {
|
|
51
|
+
async transform(chunk, _encoding, cb) {
|
|
52
|
+
index++;
|
|
53
|
+
// console.log('transform', {index})
|
|
54
|
+
// Stop processing if THROW_IMMEDIATELY mode is used
|
|
55
|
+
if (isRejected && errorMode === js_lib_1.ErrorMode.THROW_IMMEDIATELY)
|
|
56
|
+
return cb();
|
|
57
|
+
// It resolves when it is successfully STARTED execution.
|
|
58
|
+
// If it's queued instead - it'll wait and resolve only upon START.
|
|
59
|
+
await q.push(async () => {
|
|
71
60
|
try {
|
|
72
|
-
|
|
61
|
+
const currentIndex = index; // because we need to pass it to 2 functions - mapper and predicate. Refers to INPUT index (since it may return multiple outputs)
|
|
62
|
+
const res = await mapper(chunk, currentIndex);
|
|
63
|
+
const passedResults = await (0, js_lib_1.pFilter)(flattenArrayOutput && Array.isArray(res) ? res : [res], async (r) => await predicate(r, currentIndex));
|
|
64
|
+
passedResults.forEach(r => this.push(r));
|
|
73
65
|
}
|
|
74
|
-
catch {
|
|
75
|
-
|
|
76
|
-
|
|
77
|
-
|
|
78
|
-
|
|
79
|
-
|
|
80
|
-
|
|
81
|
-
|
|
82
|
-
|
|
83
|
-
|
|
84
|
-
|
|
66
|
+
catch (err) {
|
|
67
|
+
logger.error(err);
|
|
68
|
+
errors++;
|
|
69
|
+
logErrorStats(logger);
|
|
70
|
+
if (onError) {
|
|
71
|
+
try {
|
|
72
|
+
onError(err, chunk);
|
|
73
|
+
}
|
|
74
|
+
catch { }
|
|
75
|
+
}
|
|
76
|
+
if (errorMode === js_lib_1.ErrorMode.THROW_IMMEDIATELY) {
|
|
77
|
+
isRejected = true;
|
|
78
|
+
// Emit error immediately
|
|
79
|
+
// return cb(err as Error)
|
|
80
|
+
return this.emit('error', err);
|
|
81
|
+
}
|
|
82
|
+
if (errorMode === js_lib_1.ErrorMode.THROW_AGGREGATED) {
|
|
83
|
+
collectedErrors.push(err);
|
|
84
|
+
}
|
|
85
|
+
}
|
|
86
|
+
});
|
|
87
|
+
// Resolved, which means it STARTED processing
|
|
88
|
+
// This means we can take more load
|
|
85
89
|
cb();
|
|
86
|
-
}
|
|
90
|
+
},
|
|
87
91
|
});
|
|
88
92
|
function logErrorStats(logger, final = false) {
|
|
89
93
|
if (!errors)
|
|
@@ -1,7 +1,11 @@
|
|
|
1
|
-
import { AsyncMapper } from '@naturalcycles/js-lib';
|
|
1
|
+
import { AsyncMapper, Mapper } from '@naturalcycles/js-lib';
|
|
2
2
|
import { TransformMapOptions } from '../..';
|
|
3
3
|
import { WritableTyped } from '../stream.model';
|
|
4
4
|
/**
|
|
5
5
|
* Just an alias to transformMap that declares OUT as void.
|
|
6
6
|
*/
|
|
7
7
|
export declare function writableForEach<IN = any>(mapper: AsyncMapper<IN, void>, opt?: TransformMapOptions<IN, void>): WritableTyped<IN>;
|
|
8
|
+
/**
|
|
9
|
+
* Just an alias to transformMap that declares OUT as void.
|
|
10
|
+
*/
|
|
11
|
+
export declare function writableForEachSync<IN = any>(mapper: Mapper<IN, void>, opt?: TransformMapOptions<IN, void>): WritableTyped<IN>;
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
"use strict";
|
|
2
2
|
Object.defineProperty(exports, "__esModule", { value: true });
|
|
3
|
-
exports.writableForEach = void 0;
|
|
3
|
+
exports.writableForEachSync = exports.writableForEach = void 0;
|
|
4
4
|
const js_lib_1 = require("@naturalcycles/js-lib");
|
|
5
5
|
const __1 = require("../..");
|
|
6
6
|
/**
|
|
@@ -10,3 +10,10 @@ function writableForEach(mapper, opt = {}) {
|
|
|
10
10
|
return (0, __1.transformMap)(mapper, { ...opt, predicate: js_lib_1._passNothingPredicate });
|
|
11
11
|
}
|
|
12
12
|
exports.writableForEach = writableForEach;
|
|
13
|
+
/**
|
|
14
|
+
* Just an alias to transformMap that declares OUT as void.
|
|
15
|
+
*/
|
|
16
|
+
function writableForEachSync(mapper, opt = {}) {
|
|
17
|
+
return (0, __1.transformMapSync)(mapper, { ...opt, predicate: js_lib_1._passNothingPredicate });
|
|
18
|
+
}
|
|
19
|
+
exports.writableForEachSync = writableForEachSync;
|
package/package.json
CHANGED
package/src/index.ts
CHANGED
|
@@ -104,7 +104,7 @@ import {
|
|
|
104
104
|
TransformMultiThreadedOptions,
|
|
105
105
|
} from './stream/transform/worker/transformMultiThreaded'
|
|
106
106
|
import { WorkerInput, WorkerOutput } from './stream/transform/worker/transformMultiThreaded.model'
|
|
107
|
-
|
|
107
|
+
export * from './stream/writable/writableForEach'
|
|
108
108
|
import { writableFork } from './stream/writable/writableFork'
|
|
109
109
|
import { writablePushToArray } from './stream/writable/writablePushToArray'
|
|
110
110
|
import { writableVoid } from './stream/writable/writableVoid'
|
|
@@ -327,7 +327,6 @@ export {
|
|
|
327
327
|
transformMapSync,
|
|
328
328
|
transformMapSimple,
|
|
329
329
|
transformNoOp,
|
|
330
|
-
writableForEach,
|
|
331
330
|
writablePushToArray,
|
|
332
331
|
transformSplit,
|
|
333
332
|
transformToString,
|
|
@@ -1,4 +1,4 @@
|
|
|
1
|
-
import {
|
|
1
|
+
import { writablePushToArray } from '../..'
|
|
2
2
|
import { pipelineFromNDJsonFile, PipelineFromNDJsonFileOptions } from './pipelineFromNDJsonFile'
|
|
3
3
|
|
|
4
4
|
/**
|
|
@@ -9,7 +9,7 @@ export async function ndJsonFileRead<OUT = any>(
|
|
|
9
9
|
): Promise<OUT[]> {
|
|
10
10
|
const res: OUT[] = []
|
|
11
11
|
|
|
12
|
-
await pipelineFromNDJsonFile([
|
|
12
|
+
await pipelineFromNDJsonFile([writablePushToArray(res)], opt)
|
|
13
13
|
|
|
14
14
|
return res
|
|
15
15
|
}
|
|
@@ -0,0 +1,133 @@
|
|
|
1
|
+
import { Transform } from 'stream'
|
|
2
|
+
import {
|
|
3
|
+
AggregatedError,
|
|
4
|
+
AsyncMapper,
|
|
5
|
+
CommonLogger,
|
|
6
|
+
ErrorMode,
|
|
7
|
+
pFilter,
|
|
8
|
+
} from '@naturalcycles/js-lib'
|
|
9
|
+
import through2Concurrent = require('through2-concurrent')
|
|
10
|
+
import { yellow } from '../../../colors'
|
|
11
|
+
import { TransformTyped } from '../../stream.model'
|
|
12
|
+
import { TransformMapOptions } from '../transformMap'
|
|
13
|
+
|
|
14
|
+
export function notNullishPredicate(item: any): boolean {
|
|
15
|
+
return item !== undefined && item !== null
|
|
16
|
+
}
|
|
17
|
+
|
|
18
|
+
/**
|
|
19
|
+
* Like pMap, but for streams.
|
|
20
|
+
* Inspired by `through2`.
|
|
21
|
+
* Main feature is concurrency control (implemented via `through2-concurrent`) and convenient options.
|
|
22
|
+
* Using this allows native stream .pipe() to work and use backpressure.
|
|
23
|
+
*
|
|
24
|
+
* Only works in objectMode (due to through2Concurrent).
|
|
25
|
+
*
|
|
26
|
+
* Concurrency defaults to 16.
|
|
27
|
+
*
|
|
28
|
+
* If an Array is returned by `mapper` - it will be flattened and multiple results will be emitted from it. Tested by Array.isArray().
|
|
29
|
+
*/
|
|
30
|
+
export function transformMapLegacy<IN = any, OUT = IN>(
|
|
31
|
+
mapper: AsyncMapper<IN, OUT>,
|
|
32
|
+
opt: TransformMapOptions<IN, OUT> = {},
|
|
33
|
+
): TransformTyped<IN, OUT> {
|
|
34
|
+
const {
|
|
35
|
+
concurrency = 16,
|
|
36
|
+
predicate = notNullishPredicate,
|
|
37
|
+
errorMode = ErrorMode.THROW_IMMEDIATELY,
|
|
38
|
+
flattenArrayOutput,
|
|
39
|
+
onError,
|
|
40
|
+
beforeFinal,
|
|
41
|
+
metric = 'stream',
|
|
42
|
+
logger = console,
|
|
43
|
+
} = opt
|
|
44
|
+
|
|
45
|
+
let index = -1
|
|
46
|
+
let isRejected = false
|
|
47
|
+
let errors = 0
|
|
48
|
+
const collectedErrors: Error[] = [] // only used if errorMode == THROW_AGGREGATED
|
|
49
|
+
|
|
50
|
+
return through2Concurrent.obj(
|
|
51
|
+
{
|
|
52
|
+
maxConcurrency: concurrency,
|
|
53
|
+
// autoDestroy: true,
|
|
54
|
+
async final(cb) {
|
|
55
|
+
// console.log('transformMap final')
|
|
56
|
+
|
|
57
|
+
logErrorStats(logger, true)
|
|
58
|
+
|
|
59
|
+
await beforeFinal?.() // call beforeFinal if defined
|
|
60
|
+
|
|
61
|
+
if (collectedErrors.length) {
|
|
62
|
+
// emit Aggregated error
|
|
63
|
+
cb(new AggregatedError(collectedErrors))
|
|
64
|
+
} else {
|
|
65
|
+
// emit no error
|
|
66
|
+
cb()
|
|
67
|
+
}
|
|
68
|
+
},
|
|
69
|
+
},
|
|
70
|
+
async function transformMapFn(
|
|
71
|
+
this: Transform,
|
|
72
|
+
chunk: IN,
|
|
73
|
+
_encoding: any,
|
|
74
|
+
cb: (...args: any[]) => any,
|
|
75
|
+
) {
|
|
76
|
+
index++
|
|
77
|
+
// console.log({chunk, _encoding})
|
|
78
|
+
|
|
79
|
+
// Stop processing if THROW_IMMEDIATELY mode is used
|
|
80
|
+
if (isRejected && errorMode === ErrorMode.THROW_IMMEDIATELY) return cb()
|
|
81
|
+
|
|
82
|
+
try {
|
|
83
|
+
const currentIndex = index // because we need to pass it to 2 functions - mapper and predicate. Refers to INPUT index (since it may return multiple outputs)
|
|
84
|
+
const res = await mapper(chunk, currentIndex)
|
|
85
|
+
const passedResults = await pFilter(
|
|
86
|
+
flattenArrayOutput && Array.isArray(res) ? res : [res],
|
|
87
|
+
async r => await predicate(r, currentIndex),
|
|
88
|
+
)
|
|
89
|
+
|
|
90
|
+
if (passedResults.length === 0) {
|
|
91
|
+
cb() // 0 results
|
|
92
|
+
} else {
|
|
93
|
+
passedResults.forEach(r => {
|
|
94
|
+
this.push(r)
|
|
95
|
+
// cb(null, r)
|
|
96
|
+
})
|
|
97
|
+
cb() // done processing
|
|
98
|
+
}
|
|
99
|
+
} catch (err) {
|
|
100
|
+
logger.error(err)
|
|
101
|
+
|
|
102
|
+
errors++
|
|
103
|
+
|
|
104
|
+
logErrorStats(logger)
|
|
105
|
+
|
|
106
|
+
if (onError) {
|
|
107
|
+
try {
|
|
108
|
+
onError(err, chunk)
|
|
109
|
+
} catch {}
|
|
110
|
+
}
|
|
111
|
+
|
|
112
|
+
if (errorMode === ErrorMode.THROW_IMMEDIATELY) {
|
|
113
|
+
isRejected = true
|
|
114
|
+
// Emit error immediately
|
|
115
|
+
return cb(err)
|
|
116
|
+
}
|
|
117
|
+
|
|
118
|
+
if (errorMode === ErrorMode.THROW_AGGREGATED) {
|
|
119
|
+
collectedErrors.push(err as Error)
|
|
120
|
+
}
|
|
121
|
+
|
|
122
|
+
// Tell input stream that we're done processing, but emit nothing to output - not error nor result
|
|
123
|
+
cb()
|
|
124
|
+
}
|
|
125
|
+
},
|
|
126
|
+
)
|
|
127
|
+
|
|
128
|
+
function logErrorStats(logger: CommonLogger, final = false): void {
|
|
129
|
+
if (!errors) return
|
|
130
|
+
|
|
131
|
+
logger.log(`${metric} ${final ? 'final ' : ''}errors: ${yellow(errors)}`)
|
|
132
|
+
}
|
|
133
|
+
}
|
|
@@ -6,8 +6,8 @@ import {
|
|
|
6
6
|
CommonLogger,
|
|
7
7
|
ErrorMode,
|
|
8
8
|
pFilter,
|
|
9
|
+
PQueue,
|
|
9
10
|
} from '@naturalcycles/js-lib'
|
|
10
|
-
import through2Concurrent = require('through2-concurrent')
|
|
11
11
|
import { yellow } from '../../colors'
|
|
12
12
|
import { TransformTyped } from '../stream.model'
|
|
13
13
|
|
|
@@ -98,83 +98,84 @@ export function transformMap<IN = any, OUT = IN>(
|
|
|
98
98
|
let errors = 0
|
|
99
99
|
const collectedErrors: Error[] = [] // only used if errorMode == THROW_AGGREGATED
|
|
100
100
|
|
|
101
|
-
|
|
102
|
-
|
|
103
|
-
|
|
104
|
-
|
|
105
|
-
|
|
106
|
-
// console.log('transformMap final')
|
|
101
|
+
const q = new PQueue({
|
|
102
|
+
concurrency,
|
|
103
|
+
resolveOn: 'start',
|
|
104
|
+
// debug: true,
|
|
105
|
+
})
|
|
107
106
|
|
|
108
|
-
|
|
107
|
+
return new Transform({
|
|
108
|
+
objectMode: true,
|
|
109
109
|
|
|
110
|
-
|
|
110
|
+
async final(cb) {
|
|
111
|
+
// console.log('transformMap final', {index}, q.inFlight, q.queueSize)
|
|
111
112
|
|
|
112
|
-
|
|
113
|
-
|
|
114
|
-
cb(new AggregatedError(collectedErrors))
|
|
115
|
-
} else {
|
|
116
|
-
// emit no error
|
|
117
|
-
cb()
|
|
118
|
-
}
|
|
119
|
-
},
|
|
120
|
-
},
|
|
121
|
-
async function transformMapFn(
|
|
122
|
-
this: Transform,
|
|
123
|
-
chunk: IN,
|
|
124
|
-
_encoding: any,
|
|
125
|
-
cb: (...args: any[]) => any,
|
|
126
|
-
) {
|
|
127
|
-
index++
|
|
128
|
-
// console.log({chunk, _encoding})
|
|
113
|
+
// wait for the current inFlight jobs to complete and push their results
|
|
114
|
+
await q.onIdle()
|
|
129
115
|
|
|
130
|
-
|
|
131
|
-
if (isRejected && errorMode === ErrorMode.THROW_IMMEDIATELY) return cb()
|
|
116
|
+
logErrorStats(logger, true)
|
|
132
117
|
|
|
133
|
-
|
|
134
|
-
const currentIndex = index // because we need to pass it to 2 functions - mapper and predicate. Refers to INPUT index (since it may return multiple outputs)
|
|
135
|
-
const res = await mapper(chunk, currentIndex)
|
|
136
|
-
const passedResults = await pFilter(
|
|
137
|
-
flattenArrayOutput && Array.isArray(res) ? res : [res],
|
|
138
|
-
async r => await predicate(r, currentIndex),
|
|
139
|
-
)
|
|
140
|
-
|
|
141
|
-
if (passedResults.length === 0) {
|
|
142
|
-
cb() // 0 results
|
|
143
|
-
} else {
|
|
144
|
-
passedResults.forEach(r => {
|
|
145
|
-
this.push(r)
|
|
146
|
-
// cb(null, r)
|
|
147
|
-
})
|
|
148
|
-
cb() // done processing
|
|
149
|
-
}
|
|
150
|
-
} catch (err) {
|
|
151
|
-
logger.error(err)
|
|
152
|
-
|
|
153
|
-
errors++
|
|
118
|
+
await beforeFinal?.() // call beforeFinal if defined
|
|
154
119
|
|
|
155
|
-
|
|
120
|
+
if (collectedErrors.length) {
|
|
121
|
+
// emit Aggregated error
|
|
122
|
+
cb(new AggregatedError(collectedErrors))
|
|
123
|
+
} else {
|
|
124
|
+
// emit no error
|
|
125
|
+
cb()
|
|
126
|
+
}
|
|
127
|
+
},
|
|
156
128
|
|
|
157
|
-
|
|
158
|
-
|
|
159
|
-
|
|
160
|
-
} catch {}
|
|
161
|
-
}
|
|
129
|
+
async transform(this: Transform, chunk: IN, _encoding, cb) {
|
|
130
|
+
index++
|
|
131
|
+
// console.log('transform', {index})
|
|
162
132
|
|
|
163
|
-
|
|
164
|
-
|
|
165
|
-
// Emit error immediately
|
|
166
|
-
return cb(err)
|
|
167
|
-
}
|
|
133
|
+
// Stop processing if THROW_IMMEDIATELY mode is used
|
|
134
|
+
if (isRejected && errorMode === ErrorMode.THROW_IMMEDIATELY) return cb()
|
|
168
135
|
|
|
169
|
-
|
|
170
|
-
|
|
136
|
+
// It resolves when it is successfully STARTED execution.
|
|
137
|
+
// If it's queued instead - it'll wait and resolve only upon START.
|
|
138
|
+
await q.push(async () => {
|
|
139
|
+
try {
|
|
140
|
+
const currentIndex = index // because we need to pass it to 2 functions - mapper and predicate. Refers to INPUT index (since it may return multiple outputs)
|
|
141
|
+
const res = await mapper(chunk, currentIndex)
|
|
142
|
+
const passedResults = await pFilter(
|
|
143
|
+
flattenArrayOutput && Array.isArray(res) ? res : [res],
|
|
144
|
+
async r => await predicate(r, currentIndex),
|
|
145
|
+
)
|
|
146
|
+
|
|
147
|
+
passedResults.forEach(r => this.push(r))
|
|
148
|
+
} catch (err) {
|
|
149
|
+
logger.error(err)
|
|
150
|
+
|
|
151
|
+
errors++
|
|
152
|
+
|
|
153
|
+
logErrorStats(logger)
|
|
154
|
+
|
|
155
|
+
if (onError) {
|
|
156
|
+
try {
|
|
157
|
+
onError(err, chunk)
|
|
158
|
+
} catch {}
|
|
159
|
+
}
|
|
160
|
+
|
|
161
|
+
if (errorMode === ErrorMode.THROW_IMMEDIATELY) {
|
|
162
|
+
isRejected = true
|
|
163
|
+
// Emit error immediately
|
|
164
|
+
// return cb(err as Error)
|
|
165
|
+
return this.emit('error', err as Error)
|
|
166
|
+
}
|
|
167
|
+
|
|
168
|
+
if (errorMode === ErrorMode.THROW_AGGREGATED) {
|
|
169
|
+
collectedErrors.push(err as Error)
|
|
170
|
+
}
|
|
171
171
|
}
|
|
172
|
+
})
|
|
172
173
|
|
|
173
|
-
|
|
174
|
-
|
|
175
|
-
|
|
174
|
+
// Resolved, which means it STARTED processing
|
|
175
|
+
// This means we can take more load
|
|
176
|
+
cb()
|
|
176
177
|
},
|
|
177
|
-
)
|
|
178
|
+
})
|
|
178
179
|
|
|
179
180
|
function logErrorStats(logger: CommonLogger, final = false): void {
|
|
180
181
|
if (!errors) return
|
|
@@ -1,5 +1,5 @@
|
|
|
1
|
-
import { AsyncMapper, _passNothingPredicate } from '@naturalcycles/js-lib'
|
|
2
|
-
import { transformMap, TransformMapOptions } from '../..'
|
|
1
|
+
import { AsyncMapper, _passNothingPredicate, Mapper } from '@naturalcycles/js-lib'
|
|
2
|
+
import { transformMap, TransformMapOptions, transformMapSync } from '../..'
|
|
3
3
|
import { WritableTyped } from '../stream.model'
|
|
4
4
|
|
|
5
5
|
/**
|
|
@@ -11,3 +11,13 @@ export function writableForEach<IN = any>(
|
|
|
11
11
|
): WritableTyped<IN> {
|
|
12
12
|
return transformMap<IN, void>(mapper, { ...opt, predicate: _passNothingPredicate })
|
|
13
13
|
}
|
|
14
|
+
|
|
15
|
+
/**
|
|
16
|
+
* Just an alias to transformMap that declares OUT as void.
|
|
17
|
+
*/
|
|
18
|
+
export function writableForEachSync<IN = any>(
|
|
19
|
+
mapper: Mapper<IN, void>,
|
|
20
|
+
opt: TransformMapOptions<IN, void> = {},
|
|
21
|
+
): WritableTyped<IN> {
|
|
22
|
+
return transformMapSync<IN, void>(mapper, { ...opt, predicate: _passNothingPredicate })
|
|
23
|
+
}
|
|
@@ -1,98 +0,0 @@
|
|
|
1
|
-
"use strict";
|
|
2
|
-
Object.defineProperty(exports, "__esModule", { value: true });
|
|
3
|
-
exports.transformMap2 = exports.notNullishPredicate = void 0;
|
|
4
|
-
const stream_1 = require("stream");
|
|
5
|
-
const js_lib_1 = require("@naturalcycles/js-lib");
|
|
6
|
-
const colors_1 = require("../../../colors");
|
|
7
|
-
function notNullishPredicate(item) {
|
|
8
|
-
return item !== undefined && item !== null;
|
|
9
|
-
}
|
|
10
|
-
exports.notNullishPredicate = notNullishPredicate;
|
|
11
|
-
/**
|
|
12
|
-
* Like pMap, but for streams.
|
|
13
|
-
* Inspired by `through2`.
|
|
14
|
-
* Main feature is concurrency control (implemented via `through2-concurrent`) and convenient options.
|
|
15
|
-
* Using this allows native stream .pipe() to work and use backpressure.
|
|
16
|
-
*
|
|
17
|
-
* Only works in objectMode (due to through2Concurrent).
|
|
18
|
-
*
|
|
19
|
-
* Concurrency defaults to 16.
|
|
20
|
-
*
|
|
21
|
-
* If an Array is returned by `mapper` - it will be flattened and multiple results will be emitted from it. Tested by Array.isArray().
|
|
22
|
-
*/
|
|
23
|
-
function transformMap2(mapper, opt = {}) {
|
|
24
|
-
const { concurrency = 16, predicate = notNullishPredicate, errorMode = js_lib_1.ErrorMode.THROW_IMMEDIATELY, flattenArrayOutput, onError, beforeFinal, metric = 'stream', logger = console, } = opt;
|
|
25
|
-
let index = -1;
|
|
26
|
-
let isRejected = false;
|
|
27
|
-
let errors = 0;
|
|
28
|
-
const collectedErrors = []; // only used if errorMode == THROW_AGGREGATED
|
|
29
|
-
const q = new js_lib_1.PQueue({
|
|
30
|
-
concurrency,
|
|
31
|
-
resolveOn: 'start',
|
|
32
|
-
// debug: true,
|
|
33
|
-
});
|
|
34
|
-
return new stream_1.Transform({
|
|
35
|
-
objectMode: true,
|
|
36
|
-
async final(cb) {
|
|
37
|
-
// console.log('transformMap final', {index}, q.inFlight, q.queueSize)
|
|
38
|
-
// wait for the current inFlight jobs to complete and push their results
|
|
39
|
-
await q.onIdle();
|
|
40
|
-
logErrorStats(logger, true);
|
|
41
|
-
await beforeFinal?.(); // call beforeFinal if defined
|
|
42
|
-
if (collectedErrors.length) {
|
|
43
|
-
// emit Aggregated error
|
|
44
|
-
cb(new js_lib_1.AggregatedError(collectedErrors));
|
|
45
|
-
}
|
|
46
|
-
else {
|
|
47
|
-
// emit no error
|
|
48
|
-
cb();
|
|
49
|
-
}
|
|
50
|
-
},
|
|
51
|
-
async transform(chunk, _encoding, cb) {
|
|
52
|
-
index++;
|
|
53
|
-
// console.log('transform', {index})
|
|
54
|
-
// Stop processing if THROW_IMMEDIATELY mode is used
|
|
55
|
-
if (isRejected && errorMode === js_lib_1.ErrorMode.THROW_IMMEDIATELY)
|
|
56
|
-
return cb();
|
|
57
|
-
// It resolves when it is successfully STARTED execution.
|
|
58
|
-
// If it's queued instead - it'll wait and resolve only upon START.
|
|
59
|
-
await q.push(async () => {
|
|
60
|
-
try {
|
|
61
|
-
const currentIndex = index; // because we need to pass it to 2 functions - mapper and predicate. Refers to INPUT index (since it may return multiple outputs)
|
|
62
|
-
const res = await mapper(chunk, currentIndex);
|
|
63
|
-
const passedResults = await (0, js_lib_1.pFilter)(flattenArrayOutput && Array.isArray(res) ? res : [res], async (r) => await predicate(r, currentIndex));
|
|
64
|
-
passedResults.forEach(r => this.push(r));
|
|
65
|
-
}
|
|
66
|
-
catch (err) {
|
|
67
|
-
logger.error(err);
|
|
68
|
-
errors++;
|
|
69
|
-
logErrorStats(logger);
|
|
70
|
-
if (onError) {
|
|
71
|
-
try {
|
|
72
|
-
onError(err, chunk);
|
|
73
|
-
}
|
|
74
|
-
catch { }
|
|
75
|
-
}
|
|
76
|
-
if (errorMode === js_lib_1.ErrorMode.THROW_IMMEDIATELY) {
|
|
77
|
-
isRejected = true;
|
|
78
|
-
// Emit error immediately
|
|
79
|
-
// return cb(err as Error)
|
|
80
|
-
return this.emit('error', err);
|
|
81
|
-
}
|
|
82
|
-
if (errorMode === js_lib_1.ErrorMode.THROW_AGGREGATED) {
|
|
83
|
-
collectedErrors.push(err);
|
|
84
|
-
}
|
|
85
|
-
}
|
|
86
|
-
});
|
|
87
|
-
// Resolved, which means it STARTED processing
|
|
88
|
-
// This means we can take more load
|
|
89
|
-
cb();
|
|
90
|
-
},
|
|
91
|
-
});
|
|
92
|
-
function logErrorStats(logger, final = false) {
|
|
93
|
-
if (!errors)
|
|
94
|
-
return;
|
|
95
|
-
logger.log(`${metric} ${final ? 'final ' : ''}errors: ${(0, colors_1.yellow)(errors)}`);
|
|
96
|
-
}
|
|
97
|
-
}
|
|
98
|
-
exports.transformMap2 = transformMap2;
|
|
@@ -1,134 +0,0 @@
|
|
|
1
|
-
import { Transform } from 'stream'
|
|
2
|
-
import {
|
|
3
|
-
AggregatedError,
|
|
4
|
-
AsyncMapper,
|
|
5
|
-
CommonLogger,
|
|
6
|
-
ErrorMode,
|
|
7
|
-
pFilter,
|
|
8
|
-
PQueue,
|
|
9
|
-
} from '@naturalcycles/js-lib'
|
|
10
|
-
import { yellow } from '../../../colors'
|
|
11
|
-
import { TransformTyped } from '../../stream.model'
|
|
12
|
-
import { TransformMapOptions } from '../transformMap'
|
|
13
|
-
|
|
14
|
-
export function notNullishPredicate(item: any): boolean {
|
|
15
|
-
return item !== undefined && item !== null
|
|
16
|
-
}
|
|
17
|
-
|
|
18
|
-
/**
|
|
19
|
-
* Like pMap, but for streams.
|
|
20
|
-
* Inspired by `through2`.
|
|
21
|
-
* Main feature is concurrency control (implemented via `through2-concurrent`) and convenient options.
|
|
22
|
-
* Using this allows native stream .pipe() to work and use backpressure.
|
|
23
|
-
*
|
|
24
|
-
* Only works in objectMode (due to through2Concurrent).
|
|
25
|
-
*
|
|
26
|
-
* Concurrency defaults to 16.
|
|
27
|
-
*
|
|
28
|
-
* If an Array is returned by `mapper` - it will be flattened and multiple results will be emitted from it. Tested by Array.isArray().
|
|
29
|
-
*/
|
|
30
|
-
export function transformMap2<IN = any, OUT = IN>(
|
|
31
|
-
mapper: AsyncMapper<IN, OUT>,
|
|
32
|
-
opt: TransformMapOptions<IN, OUT> = {},
|
|
33
|
-
): TransformTyped<IN, OUT> {
|
|
34
|
-
const {
|
|
35
|
-
concurrency = 16,
|
|
36
|
-
predicate = notNullishPredicate,
|
|
37
|
-
errorMode = ErrorMode.THROW_IMMEDIATELY,
|
|
38
|
-
flattenArrayOutput,
|
|
39
|
-
onError,
|
|
40
|
-
beforeFinal,
|
|
41
|
-
metric = 'stream',
|
|
42
|
-
logger = console,
|
|
43
|
-
} = opt
|
|
44
|
-
|
|
45
|
-
let index = -1
|
|
46
|
-
let isRejected = false
|
|
47
|
-
let errors = 0
|
|
48
|
-
const collectedErrors: Error[] = [] // only used if errorMode == THROW_AGGREGATED
|
|
49
|
-
|
|
50
|
-
const q = new PQueue({
|
|
51
|
-
concurrency,
|
|
52
|
-
resolveOn: 'start',
|
|
53
|
-
// debug: true,
|
|
54
|
-
})
|
|
55
|
-
|
|
56
|
-
return new Transform({
|
|
57
|
-
objectMode: true,
|
|
58
|
-
|
|
59
|
-
async final(cb) {
|
|
60
|
-
// console.log('transformMap final', {index}, q.inFlight, q.queueSize)
|
|
61
|
-
|
|
62
|
-
// wait for the current inFlight jobs to complete and push their results
|
|
63
|
-
await q.onIdle()
|
|
64
|
-
|
|
65
|
-
logErrorStats(logger, true)
|
|
66
|
-
|
|
67
|
-
await beforeFinal?.() // call beforeFinal if defined
|
|
68
|
-
|
|
69
|
-
if (collectedErrors.length) {
|
|
70
|
-
// emit Aggregated error
|
|
71
|
-
cb(new AggregatedError(collectedErrors))
|
|
72
|
-
} else {
|
|
73
|
-
// emit no error
|
|
74
|
-
cb()
|
|
75
|
-
}
|
|
76
|
-
},
|
|
77
|
-
|
|
78
|
-
async transform(this: Transform, chunk: IN, _encoding, cb) {
|
|
79
|
-
index++
|
|
80
|
-
// console.log('transform', {index})
|
|
81
|
-
|
|
82
|
-
// Stop processing if THROW_IMMEDIATELY mode is used
|
|
83
|
-
if (isRejected && errorMode === ErrorMode.THROW_IMMEDIATELY) return cb()
|
|
84
|
-
|
|
85
|
-
// It resolves when it is successfully STARTED execution.
|
|
86
|
-
// If it's queued instead - it'll wait and resolve only upon START.
|
|
87
|
-
await q.push(async () => {
|
|
88
|
-
try {
|
|
89
|
-
const currentIndex = index // because we need to pass it to 2 functions - mapper and predicate. Refers to INPUT index (since it may return multiple outputs)
|
|
90
|
-
const res = await mapper(chunk, currentIndex)
|
|
91
|
-
const passedResults = await pFilter(
|
|
92
|
-
flattenArrayOutput && Array.isArray(res) ? res : [res],
|
|
93
|
-
async r => await predicate(r, currentIndex),
|
|
94
|
-
)
|
|
95
|
-
|
|
96
|
-
passedResults.forEach(r => this.push(r))
|
|
97
|
-
} catch (err) {
|
|
98
|
-
logger.error(err)
|
|
99
|
-
|
|
100
|
-
errors++
|
|
101
|
-
|
|
102
|
-
logErrorStats(logger)
|
|
103
|
-
|
|
104
|
-
if (onError) {
|
|
105
|
-
try {
|
|
106
|
-
onError(err, chunk)
|
|
107
|
-
} catch {}
|
|
108
|
-
}
|
|
109
|
-
|
|
110
|
-
if (errorMode === ErrorMode.THROW_IMMEDIATELY) {
|
|
111
|
-
isRejected = true
|
|
112
|
-
// Emit error immediately
|
|
113
|
-
// return cb(err as Error)
|
|
114
|
-
return this.emit('error', err as Error)
|
|
115
|
-
}
|
|
116
|
-
|
|
117
|
-
if (errorMode === ErrorMode.THROW_AGGREGATED) {
|
|
118
|
-
collectedErrors.push(err as Error)
|
|
119
|
-
}
|
|
120
|
-
}
|
|
121
|
-
})
|
|
122
|
-
|
|
123
|
-
// Resolved, which means it STARTED processing
|
|
124
|
-
// This means we can take more load
|
|
125
|
-
cb()
|
|
126
|
-
},
|
|
127
|
-
})
|
|
128
|
-
|
|
129
|
-
function logErrorStats(logger: CommonLogger, final = false): void {
|
|
130
|
-
if (!errors) return
|
|
131
|
-
|
|
132
|
-
logger.log(`${metric} ${final ? 'final ' : ''}errors: ${yellow(errors)}`)
|
|
133
|
-
}
|
|
134
|
-
}
|