feat: recompile for Node 10.x runtime

.circleci/config.yml

@@ -3,7 +3,7 @@ version: 2
 jobs:
   build:
     docker:
-      - image: circleci/node:8
+      - image: circleci/node:10
 
     working_directory: ~/repo
 

.gitignore

@@ -3,4 +3,4 @@ node_modules/
 *.log
 .DS_Store
 yarn.lock
-*.tar.gz
+bin/tesseract-standalone/

compile-tesseract.sh

@@ -1,20 +1,13 @@
-#!/usr/bin/env bash
+# Spin up and enter the docker container on your machine with the following command:
+# docker run -it lambci/lambda:build-nodejs10.x bash
 
-# install basic stuff required for compilation
-sudo yum-config-manager --enable epel
 
-sudo yum install -y aclocal autoconf automake cmakegcc freetype-devel gcc gcc-c++ \
+# Then run the rest of the commands inside
+
+# install basic stuff required for compilation
+yum install -y aclocal autoconf automake cmakegcc freetype-devel gcc gcc-c++ \
 	git lcms2-devel libjpeg-devel libjpeg-turbo-devel autogen autoconf libtool \
 	libpng-devel libtiff-devel libtool libwebp-devel libzip-devel make zlib-devel
-sudo yum groupinstall "Development Tools" -y
-
-# autoconf
-cd ~
-wget http://babyname.tips/mirrors/gnu/autoconf-archive/autoconf-archive-2017.09.28.tar.xz
-tar -xvf autoconf-archive-2017.09.28.tar.xz
-cd autoconf-archive-2017.09.28
-./configure && make && sudo make install
-sudo cp m4/* /usr/share/aclocal/cd ~ wget http://babynam
 
 # leptonica
 cd ~
@@ -23,29 +16,31 @@ cd leptonica/
 ./autogen.sh
 ./configure
 make
-sudo make install
+make install
 
 # tesseract
 cd ~
 git clone https://github.com/tesseract-ocr/tesseract.git
 cd tesseract
+git checkout 4.0.0
 export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig
 ./autogen.sh
 ./configure
 make
-sudo make install
+make install
 
 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 /lib64/libpng15.so.15 lib/
+cp /lib64/libtiff.so.5 lib/
+cp /lib64/libgomp.so.1 lib/
+cp /lib64/libjbig.so.2.0 lib/
 cp /usr/local/lib/liblept.so.5 lib/
 cp /usr/lib64/libjpeg.so.62 lib/
 cp /usr/lib64/libwebp.so.4 lib/
@@ -54,13 +49,16 @@ cp /usr/lib64/libstdc++.so.6 lib/
 # copy training data
 mkdir tessdata
 cd tessdata
-wget https://github.com/tesseract-ocr/tessdata_fast/raw/master/eng.traineddata
+curl -L https://github.com/tesseract-ocr/tessdata_fast/raw/master/eng.traineddata --output eng.traineddata
 
 # archive
 cd ~
-tar -zcvf tesseract.tar.gz tesseract-standalone
 
-# download from EC2 to local machine
-scp [email protected]:/home/ec2-user/tesseract.tar.gz $(pwd)
+# trim unneeded ~ 15 MB
+strip ./tesseract-standalone/**/*
+
+tar -zcvf tesseract.tar.gz tesseract-standalone
 
-# run compress-with-brotli.sh on local machine now
+# download from docker to local machine
+# 21c27dc1bf5d is docker container id, you can look it up by running "docker ps"
+docker cp 21c27dc1bf5d:/root/tesseract.tar.gz tt.tar.gz

package.json

@@ -1,7 +1,7 @@
 {
   "name": "@shelf/aws-lambda-tesseract",
   "version": "1.3.2",
-  "description": "11 MB Tesseract (with English training data) to fit inside AWS Lambda compressed with Brotli",
+  "description": "6 MB Tesseract (with English training data) to fit inside AWS Lambda",
   "license": "MIT",
   "repository": "shelfio/aws-lambda-tesseract",
   "author": {
@@ -10,7 +10,7 @@
     "url": "shelf.io"
   },
   "engines": {
-    "node": ">=8.10"
+    "node": ">=10"
   },
   "scripts": {
     "lint": "eslint . --fix",
@@ -24,12 +24,11 @@
   "keywords": [
     "lambda",
     "ocr",
-    "tesseract",
-    "brotli"
+    "tesseract"
   ],
   "dependencies": {
-    "@shelf/aws-lambda-brotli-unpacker": "0.0.2",
-    "is-image": "3.0.0"
+    "is-image": "3.0.0",
+    "tar": "4.4.10"
   },
   "devDependencies": {
     "@shelf/eslint-config": "0.3.5",

readme.md

@@ -1,6 +1,6 @@
-# aws-lambda-tesseract [![CircleCI](https://circleci.com/gh/shelfio/aws-lambda-tesseract/tree/master.svg?style=svg)](https://circleci.com/gh/shelfio/aws-lambda-tesseract/tree/master) ![](https://img.shields.io/badge/code_style-prettier-ff69b4.svg) [![Tesseract](https://img.shields.io/badge/tesserract-11_MB-brightgreen.svg)](bin/)
+# aws-lambda-tesseract [![CircleCI](https://circleci.com/gh/shelfio/aws-lambda-tesseract/tree/master.svg?style=svg)](https://circleci.com/gh/shelfio/aws-lambda-tesseract/tree/master) ![](https://img.shields.io/badge/code_style-prettier-ff69b4.svg) [![Tesseract](https://img.shields.io/badge/tesserract-6_MB-brightgreen.svg)](bin/)
 
-> 11 MB Tesseract (with English training data) to fit inside AWS Lambda compressed with Brotli
+> 6 MB Tesseract (with English training data) to fit inside AWS Lambda
 
 Inspired by [chrome-aws-lambda](https://github.com/alixaxel/chrome-aws-lambda) & [lambda-scanner-ocr](https://github.com/philippkeller/lambda-scanner-ocr)
 
@@ -10,9 +10,13 @@ Inspired by [chrome-aws-lambda](https://github.com/alixaxel/chrome-aws-lambda) &
 $ yarn add @shelf/aws-lambda-tesseract
 ```
 
+`1.x` versions of this library were compiled for Node 8.10.
+
+`2.x` was compiled for Node 10.x runtime.
+
 ## How does it work?
 
-This package contains an archive with [Tesseract 4.0 beta](https://github.com/tesseract-ocr/tesseract) compiled for usage in AWS Lambda environment.
+This package contains an archive with [Tesseract 4.0](https://github.com/tesseract-ocr/tesseract) compiled for usage in AWS Lambda environment.
 
 When a Lambda starts, it unpacks an archive with a binary to the `/tmp` folder and makes sure it's done only once per Lambda cold start.
 
@@ -38,14 +42,13 @@ unsupported by Tesseract file extensions.
 
 ## Compile It Yourself
 
-See [compile-tesseract.sh](compile-tesseract.sh) & [compress-with-brotli.sh](compress-with-brotli.sh) files
+See [compile-tesseract.sh](compile-tesseract.sh)
 
 Smoke test that it works by running `test.sh` script
 
 ## See Also
 
 - [aws-lambda-libreoffice](https://github.com/shelfio/aws-lambda-libreoffice)
-- [aws-lambda-brotli-unpacker](https://github.com/shelfio/aws-lambda-brotli-unpacker)
 - [chrome-aws-lambda](https://github.com/alixaxel/chrome-aws-lambda)
 
 ## License

src/index.js

@@ -1,21 +1,23 @@
-const {unpack} = require('@shelf/aws-lambda-brotli-unpacker');
+const {extract} = require('tar');
 const {execFileSync, execSync} = require('child_process');
 const path = require('path');
 const isImage = require('is-image');
 
 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';
+const inputPath = path.join(__dirname, '..', 'bin', 'tt.tar.gz');
+const outputPath = '/tmp/tesseract-standalone/tesseract';
 
 module.exports.getExecutablePath = async function() {
-  return unpack({inputPath, outputPath});
+  await extract({file: inputPath, cwd: '/tmp'});
+
+  return outputPath;
 };
 
 module.exports.getTextFromImage = async function(filePath) {
-  const ttBinary = await unpack({inputPath, outputPath});
+  await extract({file: inputPath, cwd: '/tmp'});
 
-  const stdout = execFileSync(ttBinary, [filePath, 'stdout', '-l', 'eng'], {
-    cwd: '/tmp/tesseract',
+  const stdout = execFileSync(outputPath, [filePath, 'stdout', '-l', 'eng'], {
+    cwd: '/tmp/tesseract-standalone',
     env: {
       LD_LIBRARY_PATH: './lib',
       TESSDATA_PREFIX: './tessdata'

test.sh

@@ -1,5 +1,9 @@
 #!/usr/bin/env bash
 
+cd bin
+tar -xvzf tt.tar.gz
+cd ..
+
 docker run --rm \
   -v "$PWD":/var/task \
-  lambci/lambda:nodejs8.10 test.handler
+  lambci/lambda:nodejs10.x test.handler