Forward Dataset to Actor or Task

# Specify the base Docker image
FROM oven/bun:1.0.7

# Next, copy the source files using the user set
# in the base image.
COPY . ./

# Install all dependencies. 
RUN bun install

# Run the image.
CMD bun run start

{
    "actorSpecification": 1,
    "name": "forward-dataset-to-actor-or-task",
    "title": "Forward dataset to Actor or task",
    "description": "Forwards specified field of dataset to Actor or task",
    "version": "0.0",
    "meta": {
        "templateId": "ts-start-bun"
    },
    "input": "./input_schema.json",
    "dockerfile": "./Dockerfile"
}

{
    "title": "Scrape data from a web page",
    "type": "object",
    "schemaVersion": 1,
    "properties": {
        "datasetId": {
            "sectionCaption": "Source dataset",
            "title": "Dataset ID",
            "type": "string",
            "description": "Id of dataset that should be forwarded",
            "editor": "textfield"
        },
        "datasetFieldName": {
            "title": "Dataset field name",
            "type": "string",
            "description": "Name of the dataset field",
            "editor": "textfield"
        },
        "targetType": {
            "sectionCaption": "Target",
            "title": "Type",
            "type": "string",
            "description": "Select if the target is Actor or task",
            "editor": "select",
            "enum": ["ACTOR", "TASK"],
            "enumTitles": ["Actor", "Task"],
            "prefill": "TASK"
        },
        "targetId": {
            "title": "Target ID or name",
            "type": "string",
            "description": "",
            "editor": "textfield"
        },
        "targetFieldName": {
            "title": "Field name",
            "type": "string",
            "description": "Name of the field on input of target Actor or task that the dataset should be mapped to.",
            "editor": "textfield"
        },
        "format": {
            "title": "Format",
            "description": "Pick the format that should be used. Corresponds to target field format.",
            "type": "string",
            "editor": "select",
            "enum": ["stringList", "requestListSources"],
            "enumTitles": ["String list", "Request list sources"],
            "prefill": "stringList"
        },
        "inputOverride": {
            "title": "Input override",
            "type": "object",
            "description": "Input override",
            "editor": "json"
        },
        "optionsOverride": {
            "title": "Options override",
            "type": "object",
            "description": "Options override",
            "editor": "json"
        }
    },
    "required": ["datasetId", "datasetFieldName", "format", "targetFieldName", "targetType", "targetId"]
}

1import { Actor, log } from 'apify';
2
3await Actor.init();
4
5interface Input {
6    datasetId: string;
7    datasetFieldName: string;
8    format: "stringList" | "requestListSources",
9    targetType: "ACTOR" | "TASK",
10    targetId: string;
11    targetFieldName: string;
12    inputOverride: object;
13    optionsOverride: object;
14}
15
16const input = await Actor.getInput<Input>();
17if (!input) throw new Error("Input is missing!");
18
19const { 
20    datasetId,
21    datasetFieldName,
22    format,
23    targetType,
24    targetId,
25    targetFieldName,
26    inputOverride = {},
27    optionsOverride = {},
28 } = input;
29
30const { apifyClient } = Actor;
31
32// Get Actor or task client
33const targetClient = targetType === 'TASK' ? apifyClient.task(targetId) : apifyClient.actor(targetId);
34// Check if the target actually exists
35if (!await targetClient.get()) {
36    await Actor.fail(`The ${targetType} "${targetId}" was not found.`);
37}
38
39// Prepare a format function, that will be applied to each item
40const formatFunction = format === 'requestListSources' ? (url) => ({url}) : (value) => value;
41
42
43// The variable that will contain the specified field from all items in dataset.
44const datasetAsArray = [];
45
46await Actor.setStatusMessage('Loading data from dataset...');
47
48// Loop over all entries in dataset, in batches.
49let offset = 0, total = 0, items = [];
50do {
51    const response = await apifyClient.dataset(datasetId).listItems({
52        fields: [ datasetFieldName ],
53        offset,
54        limit: 200,
55    });
56    items = response.items;
57    total = response.total;
58    offset += items.length;
59
60    items.forEach(sourceItem => {
61        // Format the item to desired shape
62        const targetItem = formatFunction(sourceItem[datasetFieldName]);
63        // Push it to the target array
64        datasetAsArray.push(targetItem);
65    });
66    
67    await Actor.setStatusMessage(`Loaded ${datasetAsArray.length}/${total} dataset items.`);
68} while(items.length > 0);
69
70// Update the field in input
71inputOverride[targetFieldName] = datasetAsArray;
72
73await Actor.setStatusMessage(`Dataset loaded, starting target ${targetType} ${targetId}`);
74
75// Start the target with specified input and options, don't wait for finish.
76await targetClient.start(inputOverride, { ...optionsOverride, waitForFinish: 0});
77
78await Actor.exit();

# configurations
.idea

# crawlee and apify storage folders
apify_storage
crawlee_storage
storage

# installed files
node_modules

# git folder
.git

# dist folder
dist

storage
apify_storage
crawlee_storage
node_modules
dist
tsconfig.tsbuildinfo
storage/*
!storage/key_value_stores
storage/key_value_stores/*
!storage/key_value_stores/default
storage/key_value_stores/default/*
!storage/key_value_stores/default/INPUT.json

{
    "name": "ts-start-bun",
    "version": "0.0.1",
    "type": "module",
    "description": "This is an example of an Apify actor.",
    "engines": {
        "bun": ">=1.0.0"
    },
    "dependencies": {
        "apify": "^3.1.10"
    },
    "devDependencies": {
        "@apify/tsconfig": "^0.1.0"
    },
    "scripts": {
        "start": "bun src/main.ts",
        "test": "echo \"Error: oops, the actor has no tests yet, sad!\" && exit 1"
    },
    "author": "It's not you it's me",
    "license": "ISC"
}

{
    "extends": "@apify/tsconfig",
    "compilerOptions": {
        "module": "ES2022",
        "target": "ES2022",
        "outDir": "dist",
        "noUnusedLocals": false,
        "lib": ["DOM"]
    },
    "include": [
        "./src/**/*"
    ]
}