Local installation of Grew-match

Grew-match is available online on a set of corpora (mainly from the UD project). If you want to use Grew-match on your own corpus, you have to install it locally, following the instructions on this page.

STEP 0: Run a web server

A web server is required. You can install apache or one of the easy to install distribution like LAMP on Linux or MAMP on Mac OSX.

In the following, we will call DOCUMENT_ROOT the main folder accessible from your website:

In needed, refer to the documentation of the corresponding web server.

We use here the port number 8888. You may have to change this if this port number is already used.

STEP 1: Install the webpage


The code for the webpage is available on gitlab.inria.fr:

git clone https://gitlab.inria.fr/grew/grew_match.git


Move to the main folder of the project:

cd grew_match

Edit the file corpora/groups.json to describe the set of available corpora. An example of configuration file with 3 corpora:

{ "groups": [
    { "id": "local",
      "name": "Local corpora",
      "corpora": [
        { "id": "my_corpora" },
        { "folder": "Older versions",
          "corpora": [
            { "id": "my_corpora@2.0" },
            { "id": "my_corpora@1.0" }

In JSON, groups defines the items in the top navbar and corpora the list of corpora in the left bar, maybe organised in folders (recursive folders are not handled). You can look the configuration file used on Grew-match for a larger example.


The project contains a file install_template.sh. Copy it with the name install.sh:

cp install_template.sh install.sh

Edit the new file install.sh and update DEST definition (line 2) and PORT (line 5) if needed. Run the install script:


STEP 2: Install the daemon

You have to start locally a daemon which will handle your requests on your corpora.


Follow general instruction for Grew installation and then install the daemon with:

opam install grew_daemon


To configure your daemon, you have to describe the corpora you want to use in a conf file. This file describes each corpora with a name, a directory and a list of files. For instance, the JSON file my_corpora.json below defines 3 corpora:

{ "corpora": [
  { "id": "my_corpora",
    "directory": "/users/me/corpora/my_corpora",
    "files": [ "my_corpora_dev.conll", "my_corpora_test.conll", "my_corpora_train.conll" ]
  { "id": "my_corpora@2.0",
    "directory": "/users/me/corpora/my_corpora/2.0",
    "files": [ "my_corpora_dev.conll", "my_corpora_test.conll", "my_corpora_train.conll" ]
  { "id": "my_corpora@1.0",
    "directory": "/users/me/corpora/my_corpora/1.0",
    "files": [ "my_corpora_dev.conll", "my_corpora_test.conll", "my_corpora_train.conll" ]

Compile your corpora

In order to speed up the pattern search and to preserve memory when a large number of corpora are available, corpora are compiled.

The compilation is done with the command:

grew compile -grew_match_server DOCUMENT_ROOT -i my_corpora.json

A new file with the name of the corpus and the extension .marshal is created in the corpus directory. Of course, you will have to compile again if one of your corpora is modified. The compilation step will also build the relation tables and put them in a place where they can be found by the server.

You can clean the compiled files with:

grew clean -i my_corpora.json

Run the daemon

The Daemon is started with the command (update the port number if necessary):

grew_daemon run --port 8888 my_corpora.json

Step 3 and more


Make sure that the web server is running. You should be able to request your corpora from http://localhost:8888/grew_match. Feel free to contact us in case of trouble.

Modification of the configuration file

If the configuration file (like my_corpora.json above) is modified, you should run again install.sh.

Restart the daemon when one of the corpora is updated

  1. Kill the running daemon (you can use the command killall grew_daemon if the daemon is running in the background)
  2. Run the compile operation again: grew compile -grew_match_server DOCUMENT_ROOT -i my_corpora.json
  3. Restart the daemon: grew_daemon run --port 8888 my_corpora.json