Skip to content

yardstickpics/tools

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

35 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Yardstick image test suite tools

Library that helps using and managing image metadata provided by Yardstick Pictures.

Requires Node.js 7.6 (for older Node.js install yr@3) and (optionally) ImageMagick. To get started:

npm install --save yr
git clone https://github.com/yardstickpics/metadata.git

download.js

Scans metadata/ and downloads all associated image files to downloads/ directory. This will download about 80GB of data.

import.js

Scans metadata/ and downloads/ directories to import all images into an sqlite database images.db.

Metadata

Programmatic API for using/browsing the metadata. The class' constructor takes following options:

  • sha1s — optional array of SHA-1 hashes of images to interate. If this option is not provided, it iterates over all available images.
  • root — optional path of clone of the metadata. ./metadata/ by default.
const yr = require('yr');
const metadata = new yr.Metadata();

The class has following methods:

.map([options,] callback)

Runs callback on every image's metadata. The callback gets an Image instance as an argument, and may return a Promise to perform asynchronous tasks.

Options are:

  • progress — boolean. If true, periodically log how many callbacks have been executed and how much time is estimated to finish.
  • cpus — integer. Number of tasks to run in parallel.
  • max — integer. Iterate only over this many images. Useful for testing tools on small samples.
  • ignoreErrors — boolean. By default if any callback throws or returns a Promise that fails, the whole map will be aborted.

Returns a Promise for an array of results from all callbacks.

Example

Iterate over all available images and file sizes of some of them:

const fs = require('fs');
const Metadata = require('yr').Metadata;
const yr = new Metadata();

yr.map({progress: true}, image => {
    if (image.data.lic == "pd") {
        return fs.statSync(image.sourcePath()).size;
    }
}).then(allValues => {
    const sizes = allValues.filter(x => x); // Remove undefined values
    const sum = sizes.reduce((sum,x) => sum+x, 0);
    console.log("Average public domain image size in this set is", sum / sizes.length);
});

.forEach([options,] callback)

Same as map, but returns a Promise for undefined.

Example

Iterate over all available images' metadata and put in any .name fields that may be missing:

const Metadata = require('yr').Metadata;
const yr = new Metadata();

yr.forEach(image => {
    if (!image.data.name) {
        image.data.name = "Unnamed image";
        image.save();
    }
})
.catch(err => console.error(err));

Image

Object that represents an image (its metadata). It has following properties/methods:

.data

Raw metadata.

.sourcePath()

Returns file path to the source image (e.g. "downloads/00/0000…0000.jpeg"). It'll return a path even if the file does not exist.

.addTag(tag)

Adds a tag to .data.tags and returns true if it's a new tag.

.json()

Returns stringified .data.

.save()

Writes .data to disk.

Image.createFromFile(metadataObject, filePath)

A "static" method on Image that creates a new metadata JSON file and copies the file to its intended location in downloads/ directory.

The first argument must be an object that contains as many fields of Raw metadata as possible, except the sha1 field, which will be set automatically.

Returns a Promise for the Image instance.

Image.createFromBuffer(metadataObject, Buffer)

Same as createFromFile, but writes image from a Buffer instead.

const Image = require('yr').Image;

yr.Image.createFromFile({lic:'pd', name:'example'}, 'example.png')
    .catch(err => console.error);