Replace URL parser with RegEx with configurable path prefix

Partial fix for #samvera/serverless-iiif#52 and #samvera/serverless-iiif#73

CHANGELOG.md

@@ -2,6 +2,10 @@
 
 Only features and major fixes are listed. Everything else can be considered a minor bugfix or maintenance release.
 
+##### v3.0.0
+
+- Add `pathPrefix` option (default: `/iiif/2/`) to constructor instead of popping a specific number of path segments off of the end of the URL
+
 ##### v2.0.0
 
 - Pass `baseUrl` to `streamResolver` and `dimension` functions

README.md

@@ -28,6 +28,7 @@ const processor = new IIIF.Processor(url, streamResolver, opts);
   * `density` (integer) – the pixel density to be included in the result image in pixels per inch
     * This has no effect whatsoever on the size of the image that gets returned; it's simply for convenience when using
       the resulting image in software that calculates a default print size based on the height, width, and density
+  * `pathPrefix` (string) – the default prefix that precedes the `id` part of the URL path (default: `/iiif/2/`)
 
 ## Examples
 

index.js

@@ -3,39 +3,19 @@ const mime = require('mime-types');
 const transform = require('./lib/transform');
 const IIIFError = require('./lib/error');
 
-const filenameRe = /(color|gray|bitonal|default)\.(jpe?g|tiff?|gif|png|webp)/;
-
-function parseUrl (url) {
-  const result = {};
-  const segments = url.split('/');
-  result.filename = segments.pop();
-  if (result.filename.match(filenameRe)) {
-    result.rotation = segments.pop();
-    result.size = segments.pop();
-    result.region = segments.pop();
-    result.quality = RegExp.$1;
-    result.format = RegExp.$2;
-  }
-  result.id = decodeURIComponent(segments.pop());
-  result.baseUrl = segments.join('/');
-  return result;
-}
+const DefaultPathPrefix = '/iiif/2/';
 
 class Processor {
   constructor (url, streamResolver, ...args) {
     const opts = this.parseOpts(args);
 
-    this
-      .initialize(url, streamResolver)
-      .setOpts(opts);
-
-    if (!filenameRe.test(this.filename) && this.filename !== 'info.json') {
-      throw new IIIFError(`Invalid IIIF URL: ${url}`);
-    }
-
     if (typeof streamResolver !== 'function') {
       throw new IIIFError('streamResolver option must be specified');
     }
+
+    this
+      .setOpts(opts)
+      .initialize(url, streamResolver);
   }
 
   parseOpts (args) {
@@ -59,19 +39,39 @@ class Processor {
     this.maxWidth = opts.maxWidth;
     this.includeMetadata = !!opts.includeMetadata;
     this.density = opts.density || null;
+    this.pathPrefix = opts.pathPrefix?.replace(/^\/*/, '/').replace(/\/*$/, '/') || DefaultPathPrefix;
+
     return this;
   }
 
-  initialize (url, streamResolver) {
-    let params = url;
+  parseUrl (url) {
+    let result;
+
     if (typeof url === 'string') {
-      params = parseUrl(params);
+      const parser = new RegExp(`(?<baseUrl>https?://[^/]+${this.pathPrefix})(?<path>.+)$`);
+      const { baseUrl, path } = parser.exec(url).groups;
+      result = transform.IIIFRegExp.exec(path)?.groups;
+      if (result === undefined) {
+        throw new IIIFError(`Invalid IIIF URL: ${url}`);
+      }
+      result.baseUrl = baseUrl;
+    } else {
+      result = url;
     }
+
+    return result;
+  }
+
+  initialize (url, streamResolver) {
+    const params = this.parseUrl(url);
+
     Object.assign(this, params);
     this.streamResolver = streamResolver;
 
     if (this.quality && this.format) {
       this.filename = [this.quality, this.format].join('.');
+    } else if (this.info) {
+      this.filename = 'info.json';
     }
     return this;
   }

lib/transform.js

@@ -19,11 +19,12 @@ function validator (type) {
   if (result instanceof Array) {
     result = result.join('|');
   }
-  return new RegExp('^(' + result + ')$');
+  return `(?<${type}>${result})`;
 }
 
 function validate (type, v) {
-  if (!validator(type).test(v)) {
+  const re = new RegExp(`^${validator(type)}$`);
+  if (!re.test(v)) {
     throw new IIIFError(`Invalid ${type}: ${v}`);
   }
   return true;
@@ -36,7 +37,15 @@ function validateDensity (v) {
     throw new IIIFError(`Invalid density value: ${v}`);
   }
   return true;
-};
+}
+
+function iiifRegExp () {
+  const transformation =
+    ['region', 'size', 'rotation'].map(type => validator(type)).join('/') +
+    '/' + validator('quality') + '.' + validator('format');
+
+  return new RegExp(`^/?(?<id>.+?)/(?:(?<info>info.json)|${transformation})$`);
+}
 
 class Operations {
   constructor (dims) {
@@ -207,6 +216,7 @@ class Operations {
 module.exports = {
   Qualities: Validators.quality,
   Formats: Validators.format,
+  IIIFRegExp: iiifRegExp(),
   Operations,
   IIIFError
 };

package.json

@@ -1,6 +1,6 @@
 {
   "name": "iiif-processor",
-  "version": "2.0.1",
+  "version": "3.0.0",
   "description": "IIIF 2.1 Image API modules for NodeJS",
   "main": "index.js",
   "repository": "https://github.com/samvera-labs/node-iiif",

tests/integration.test.js

@@ -13,7 +13,7 @@ let consoleWarnMock;
 
 describe('info.json', () => {
   beforeEach(() => {
-    subject = new iiif.Processor(`${base}/info.json`, streamResolver);
+    subject = new iiif.Processor(`${base}/info.json`, streamResolver, { pathPrefix: 'iiif/2/ab/cd/ef/gh' });
   });
 
   it('produces a valid info.json', async () => {

tests/processor.test.js

@@ -14,6 +14,8 @@ describe('IIIF Processor', () => {
   });
 
   it('Parse URL', () => {
+    assert.strictEqual(subject.id, 'ab/cd/ef/gh/i');
+    assert.strictEqual(subject.baseUrl, 'https://example.org/iiif/2/');
     assert.strictEqual(subject.rotation, '45');
     assert.strictEqual(subject.size, 'pct:50');
     assert.strictEqual(subject.region, '10,20,30,40');
@@ -166,14 +168,14 @@ describe('stream processor', () => {
 
     const streamResolver = ({ id, baseUrl }) => {
       expect(id).toEqual('i');
-      expect(baseUrl).toEqual('https://example.org/iiif/2/ab/cd/ef/gh');
+      expect(baseUrl).toEqual('https://example.org/iiif/2/ab/cd/ef/gh/');
 
       return new Stream.Readable({
         read() {}
       });
     }
 
-    const subject = new iiif.Processor(`https://example.org/iiif/2/ab/cd/ef/gh/i/10,20,30,40/pct:50/45/default.png`, streamResolver);
+    const subject = new iiif.Processor(`https://example.org/iiif/2/ab/cd/ef/gh/i/10,20,30,40/pct:50/45/default.png`, streamResolver, {pathPrefix: 'iiif/2/ab/cd/ef/gh'});
     subject.execute();
   })
 })
@@ -190,14 +192,14 @@ describe('dimension function', () => {
 
     const dimensionFunction = ({ id, baseUrl }) => {
       expect(id).toEqual('i');
-      expect(baseUrl).toEqual('https://example.org/iiif/2/ab/cd/ef/gh');
+      expect(baseUrl).toEqual('https://example.org/iiif/2/ab/cd/ef/gh/');
       return { w: 100, h: 100 }
     }
 
     const subject = new iiif.Processor(
       `https://example.org/iiif/2/ab/cd/ef/gh/i/10,20,30,40/pct:50/45/default.png`,
       streamResolver,
-      { dimensionFunction }
+      { dimensionFunction, pathPrefix: 'iiif/2/ab/cd/ef/gh' }
     );
     subject.execute();
   })