Enhancement docker build & words and bounds results

.dockerignore

@@ -0,0 +1,11 @@
+*.swo
+*.swp
+.git
+.DS_Store
+node_modules
+bin
+build
+scripts
+coverage
+.nyc_output
+.

.gitignore

@@ -4,3 +4,4 @@ node_modules/
 .DS_Store
 yarn.lock
 *.tar.gz
+build

Dockerfile

@@ -0,0 +1,16 @@
+from lambci/lambda:build-nodejs8.10
+
+RUN yum install -y autoconf aclocal automake install libtool libjpeg-devel \
+	libpng-devel libtiff-devel zlib-devel wget gzip make cmakegcc freetype-devel \
+  gcc gcc-c++ git lcms2-devel libjpeg-turbo-devel autogen libpng-devel \
+  libtiff-devel libwebp-devel libzip-devel zlib-devel libgcc
+
+RUN yum groupinstall "Development Tools" -y
+
+RUN yum install -y cmake
+
+COPY . .
+
+RUN wget https://github.com/google/brotli/archive/v1.0.7.tar.gz
+RUN tar -zxvf v1.0.7.tar.gz
+RUN cd brotli-1.0.7 && mkdir out && cd out && ../configure-cmake && make && make test && make install

Makefile

@@ -0,0 +1,18 @@
+NAME=aws-lambda-tesseract
+PWD=$(shell pwd)
+
+build: build-docker build-tesseract compress-tesseract
+
+build-dist: build-tesseract compress-tesseract
+
+build-docker:
+	docker build -f Dockerfile -t $(NAME) .
+
+version:
+	@echo $(shell git rev-parse HEAD)
+
+build-tesseract:
+	docker run -it -v $(PWD)/scripts:/scripts -v $(PWD)/build:/build -v $(PWD)/build:/build $(NAME) /scripts/compile-tesseract.sh
+
+compress-tesseract:
+	docker run -it -v $(PWD)/scripts:/scripts -v $(PWD)/build:/build $(NAME) /scripts/compress-with-brotli.sh

package.json

@@ -29,6 +29,7 @@
   ],
   "dependencies": {
     "@shelf/aws-lambda-brotli-unpacker": "0.0.2",
+    "csv-parse": "^4.3.0",
     "is-image": "2.0.0"
   },
   "devDependencies": {

readme.md

@@ -18,18 +18,49 @@ When a Lambda starts, it unpacks an archive with a binary to the `/tmp` folder a
 
 ## Usage
 
+#### Using a path
+
 ```js
 const {getTextFromImage, isSupportedFile} = require('@shelf/aws-lambda-tesseract');
 
 module.exports.handler = async event => {
   // assuming there is a photo.jpg inside /tmp dir
-  // original file will be deleted afterwards
-
   if (!isSupportedFile('/tmp/photo.jpg')) {
     return false;
   }
 
-  return getTextFromImage('/tmp/photo.jpg');
+  getTextFromImage('/tmp/photo.jpg').then(result => console.log(result));
+};
+```
+
+#### Using a stream
+
+This is useful for when you want to stream the file data from a remote source like a URL.
+
+```js
+const https = require('https');
+const {getTextFromImage, isSupportedFile} = require('@shelf/aws-lambda-tesseract');
+
+module.exports.handler = async event => {
+  // assuming that the url exists and is readable.
+  const url = 'https://cdn-std.dprcdn.net/files/acc_55602/9X4IIL';
+  const fileStream = await new Promise(resolve => https.get(url, resolve));
+  getTextFromImage(fileStream).then(result => console.log(result));
+};
+```
+
+#### Extracting words and their coordinates
+
+The `getWordsAndBounds` function returns a JSON object of extracted words and their coordinates on the page.
+
+```js
+const {getWordsAndBounds} = require('@shelf/aws-lambda-tesseract');
+
+module.exports.handler = async event => {
+  // assuming that photo.jpg exists and is readable.
+  const file = fs.createReadStream(__dirname + '/photo.jpg');
+
+  getWordsAndBounds(file).then(result => console.log(result));
 };
 ```
 
@@ -38,7 +69,15 @@ unsupported by Tesseract file extensions.
 
 ## Compile It Yourself
 
-See [compile-tesseract.sh](compile-tesseract.sh) & [compress-with-brotli.sh](compress-with-brotli.sh) files
+Compile Tesseract for deployment on Lambda. Requires [Docker](https://www.docker.com/) & [Make](https://www.gnu.org/software/make/manual/html_node/Introduction.html) to be installed.
+
+`$ make build`: Builds Docker image, compiles Tesseract 4.0.0, and compresses result into the `tt.tar.br` archive.
+
+`$ make build-tesseract`: Compiles Tesseract 4.0.0 and creates `tesseract.tar.gz` file as output.
+
+`$ make compress-tesseract`: Runs brotli compression on built Tesseract and compresses `tesseract.tar.gz` into `tt.tar.bz`.
+
+**Note:** After compiling and compressing you need to copy the latest `tt.tar.bz` into the `/bin` directory. `$ cp ./build/tt.tar.bz ./bin`
 
 ## See Also
 

scripts/compile-tesseract.sh

@@ -0,0 +1,61 @@
+#!/usr/bin/env bash
+
+echo "Building"
+# Build leptonica
+wget http://www.leptonica.com/source/leptonica-1.77.0.tar.gz
+tar -zxvf leptonica-1.77.0.tar.gz
+ls -la ./
+cd leptonica-1.77.0
+ls -la 
+./configure
+make
+make install
+
+# Build tesseract 4.0
+cd ..
+wget https://github.com/tesseract-ocr/tesseract/archive/4.0.0.tar.gz
+tar -zxvf 4.0.0.tar.gz
+cd tesseract-4.0.0/
+export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig
+./autogen.sh
+./configure
+make
+make install
+ldconfig
+
+cd ~
+mkdir tesseract-standalone
+
+# trim unneeded ~ 15 MB
+strip ./tesseract-standalone/**/*
+
+# copy files
+cd tesseract-standalone
+cp /usr/local/bin/tesseract .
+mkdir lib
+cp /usr/local/lib/libtesseract.so.4 lib/
+cp /usr/local/lib/liblept.so.5 lib/
+# cp /usr/lib64/* lib/
+cp /usr/lib64/libjpeg.so.62 lib/
+cp /usr/lib64/libwebp.so.4 lib/
+cp /usr/lib64/libstdc++.so.6 lib/
+cp /usr/lib64/libpng15.so.15 lib/
+cp /usr/lib64/libtiff.so.5 lib/
+cp /usr/lib64/libgomp.so.1 lib/
+cp /usr/lib64/libjbig.so.2.0 lib/
+
+# copy training data
+mkdir tessdata
+cd tessdata
+wget https://github.com/tesseract-ocr/tessdata_fast/raw/master/eng.traineddata
+
+# Create configs
+mkdir configs
+echo "tessedit_create_tsv 1" > configs/tsv
+
+# archive
+cd ~
+tar -zcvf tesseract.tar.gz tesseract-standalone
+mv tesseract.tar.gz /build/
+
+echo "Done!"

scripts/compress-brotli.sh

@@ -0,0 +1,10 @@
+#!/usr/bin/env bash
+
+tar -C /build -zxvf /build/tesseract.tar.gz
+mv /build/tesseract-standalone /build/tesseract
+cd /build
+tar -cf tt.tar tesseract
+rm -rf tesseract
+echo "Running brotli (this can take a few minutes)"
+brotli --best --force --verbose /build/tt.tar
+echo "Done"

src/index.js

@@ -1,30 +1,71 @@
-const {unpack} = require('@shelf/aws-lambda-brotli-unpacker');
-const {execFileSync, execSync} = require('child_process');
+const {execFile} = require('child_process');
 const path = require('path');
+const fs = require('fs');
+const parseCSV = require('csv-parse');
 const isImage = require('is-image');
+const {unpack} = require('@shelf/aws-lambda-brotli-unpacker');
 
 const unsupportedExtensions = new Set(['ai', 'emf', 'eps', 'gif', 'ico', 'psd', 'svg']);
 const inputPath = path.join(__dirname, '..', 'bin', 'tt.tar.br');
 const outputPath = '/tmp/tesseract/tesseract';
 
+async function runTesseract(file, opts) {
+  const ttBinary = process.env.TESSERACT_BINARY_PATH || (await unpack({inputPath, outputPath}));
+  let processFile = 'stdin';
+  if (typeof file === 'string' && fs.existsSync(file)) processFile = file;
+  if (!file) processFile = false;
+
+  const options = {
+    env: {}
+  };
+  if (!process.env.TESSERACT_BINARY_PATH) {
+    options.env.LD_LIBRARY_PATH =
+      `${process.env.LD_LIBRARY_PATH}:/tmp/tesseract/lib` || `/tmp/tesseract/lib`;
+    options.env.TESSDATA_PREFIX = process.env.TESSDATA_PREFIX || `/tmp/tesseract/tessdata`;
+  }
+  if (!process.env.TESSERACT_BINARY_PATH) options.cwd = '/tmp/tesseract';
+  return new Promise((resolve, reject) => {
+    const finalOpts = processFile ? [processFile, ...opts] : opts;
+    const child = execFile(ttBinary, finalOpts, options, (error, stdout, stderr) => {
+      if (error) return reject(error);
+      return resolve(stdout);
+    });
+    if (processFile === 'stdin') file.pipe(child.stdin);
+  });
+}
+
+function isUnsupportedFileExtension(filePath) {
+  const ext = path
+    .extname(filePath)
+    .slice(1)
+    .toLowerCase();
+
+  return unsupportedExtensions.has(ext);
+}
+
 module.exports.getExecutablePath = async function() {
   return unpack({inputPath, outputPath});
 };
 
-module.exports.getTextFromImage = async function(filePath) {
-  const ttBinary = await unpack({inputPath, outputPath});
-
-  const stdout = execFileSync(ttBinary, [filePath, 'stdout', '-l', 'eng'], {
-    cwd: '/tmp/tesseract',
-    env: {
-      LD_LIBRARY_PATH: './lib',
-      TESSDATA_PREFIX: './tessdata'
-    }
-  });
+module.exports.getTextFromImage = async function(file) {
+  const result = await runTesseract(file, ['stdout', '-l', 'eng']);
+  return result.toString();
+};
 
-  execSync(`rm ${filePath}`);
+module.exports.getWordsAndBounds = async function(file) {
+  const result = await runTesseract(file, ['stdout', '-l', 'eng', 'tsv']);
+  const object = await new Promise((resolve, reject) =>
+    parseCSV(result.toString(), {delimiter: '\t', columns: true}, (err, result) => {
+      if (err) return reject(err);
+      return resolve(result);
+    })
+  );
+  return object;
+};
 
-  return stdout.toString();
+module.exports.version = async function() {
+  const result = await runTesseract(false, ['--version']);
+  return result.toString();
 };
 
 module.exports.isSupportedFile = function(filePath) {
@@ -35,12 +76,3 @@ module.exports.isSupportedFile = function(filePath) {
 
   return !isUnsupportedFileExtension(filePath);
 };
-
-function isUnsupportedFileExtension(filePath) {
-  const ext = path
-    .extname(filePath)
-    .slice(1)
-    .toLowerCase();
-
-  return unsupportedExtensions.has(ext);
-}