A Go wrapper library to convert PDF, DOC, DOCX, XML, HTML, RTF, ODT, Pages documents and images (see optional dependencies below) to plain text.

Note for returning users: the Go import path for this package changed to code.sajari.com/docconv.

If you haven't setup Go before, you first need to install Go.

To fetch and build the code:

$ go install code.sajari.com/docconv/docd@latest

See go help install for details on the installation location of the installed docd executable. Make sure that the full path to the executable is in your PATH environment variable.

tidy, wv, popplerutils, unrtf, https://github.com/JalfResi/justext

Example install of dependencies (not all systems):

$ sudo apt-get install poppler-utils wv unrtf tidy
$ go get github.com/JalfResi/justext

To add image support to the docconv library you first need to install and build gosseract.

Now you can add -tags ocr to any go command when building/fetching/testing docconv to include support for processing images:

$ go get -tags ocr code.sajari.com/docconv/...

This may complain on macOS, which you can fix by installing tesseract via brew:

$ brew install tesseract

$ docker build --no-cache -t bomayu:docd -f docd/Dockerfile .

The docd tool runs as either:

1. a service on port 8888 (by default)

   Documents can be sent as a multipart POST request and the plain text (body) and meta information are then returned as a JSON object.

2. a service exposed from within a Docker container

   This also runs as a service, but from within a Docker container. Official images are published at https://hub.docker.com/r/sajari/docd.

   Optionally you can build it yourself:

   cd docd
   docker build -t docd .
   

3. via the command line.

   Documents can be sent as an argument, e.g.

   $ docd -input document.pdf
   

• addr - the bind address for the HTTP server, default is ":8888"
• log-level
  • 0: errors & critical info
  • 1: inclues 0 and logs each request as well
  • 2: include 1 and logs the response payloads
• readability-length-low - sets the readability length low if the ?readability=1 parameter is set
• readability-length-high - sets the readability length high if the ?readability=1 parameter is set
• readability-stopwords-low - sets the readability stopwords low if the ?readability=1 parameter is set
• readability-stopwords-high - sets the readability stopwords high if the ?readability=1 parameter is set
• readability-max-link-density - sets the readability max link density if the ?readability=1 parameter is set
• readability-max-heading-distance - sets the readability max heading distance if the ?readability=1 parameter is set
• readability-use-classes - comma separated list of readability classes to use if the ?readability=1 parameter is set

$ # This will only log errors and critical info
$ docd -log-level 0

$ # This will run on port 8000 and log each request
$ docd -addr :8000 -log-level 1

Some basic code is shown below, but normally you would accept the file by HTTP or open it from the file system.

This should be enough to get you started though.

Note: this assumes you have the dependencies installed.

package main

import (
	"fmt"
	"log"

	"code.sajari.com/docconv"
)

func main() {
	res, err := docconv.ConvertPath("your-file.pdf")
	if err != nil {
		log.Fatal(err)
	}
	fmt.Println(res)
}

package main

import (
	"fmt"
	"log"

	"code.sajari.com/docconv/client"
)

func main() {
	// Create a new client, using the default endpoint (localhost:8888)
	c := client.New()

	res, err := client.ConvertPath(c, "your-file.pdf")
	if err != nil {
		log.Fatal(err)
	}
	fmt.Println(res)
}

Alternatively, via a curl:

curl -s -F input=your-file.pdf https://localhost:8888/convert