Introduction

This web-server for running the i2p-website is a collection of scripts (aka The Python Scripts) and content-files to:

• manage updates (based on git),
• manage translations (generating translation files before the web-server is run),
• manage tags (generating tag files before the web-server is run),
• run a web-server creating/delivering pages on-demand (using WSGI)

This is not a static web-site generator. To see the pages you will need to setup your system for the python and shell-scripts and run the web-server contained as described. Due to heavy use of tags even content changes quickly will require a build environment to check your changes (towards breaking the build process). Translations can be done using solely a web-site and then do not require any of this (others will integrate all changes from the web-site using these scripts).

The authors are the I2P team. For details about licensing see LICENSE.txt.

Requirements Overview

If you don't want to deal with the requirements/software, you can use a docker config (see Dockerfile) which will set these up automatically. Otherwise you will need to satisfy the following requirements (that Dockerfile contains the commands for Debian btw.):

• git
• python2
• pip
• virtualenv
• apache (using WSGI to call the scripts)
• ctags? (was mentioned to be needed as both, system package + python package, but it seems only the python package is being installed?)
• transifex-client? (There is a transifex-client in Debian which might be needed for the translation steps described below?)

Note that the scripts will install additional software packages (see /etc/reqs.txt) from outside your distribution (into the virtual environment if using docker) using pip and then do some custom patching (meaning pinned versions?).

Note also that the manual way described in the following suggests to use proxychains with Tor to avoid Clearnet traffic, while the Docker version seems to use Clearnet for that.

I2P website

To run locally (for testing purposes):

• Install virtualenv and Python 2.7

• (Optional) Install proxychains, configure it for Tor

• Pull in the dependencies:

  $ proxychains ./setup_venv.sh
  

  (you can also pull them non-anon by leaving out proxychains)

• Compile translations (if you want to see them):

  $ ./compile-messages.sh
  

• Start the webserver:

  $ source env/bin/activate # activates virtualenv
  $ ./runserver.py
  $ deactivate # ..s virtualenv
  

  (if the shell in use is not bash, you can append its name to the activator if supported: ...ivate.fish)

• Open the site at http://localhost:5000/

Running a mirror

If you want to mirror the I2P website, thanks! Here is a checklist:

• Do not edit any of the files in i2p2www/
  • In particular, do not change the CANONICAL_DOMAIN variable in i2p2www/__init__.py, it needs to point to the official site for SEO.
• If you need to edit variables in etc/update.vars, copy the file to etc/update.vars.custom and edit appropriately.
• If you want to enable caching, copy i2p2www/settings.py.sample to i2p2www/settings.py and edit appropriately.
• Add ./site-updater.sh to your crontab. This will keep the site updated, recompile the translations when necessary, and touch a file in /tmp/ (look in etc/update.vars for the filename, your webserver should restart WSGI when the timestamp of this file changes).

Running a mirror with Docker

It's possible to set up a mirror using apache2 inside of a Docker container. It is intended to provide a HTTP-only server. To use HTTPS, using a reverse proxy is the easiest way. You should not need to make any modifications to the service running inside the container, but you may make the same modifications to the containerized mirror that you would to a normal mirror by changing your local copy of the site according to the recommendations in the previous settings.

• To automatically start an HTTP mirror on port 8090, run: site-updater-docker.sh

• When you have your mirror configured, add site-updater-docker.sh to your crontab to keep the site up-to-date.

Configuration and Translations

Configuration files for the various scripts are in etc/. Environment variables in etc/translation.vars can be overridden by creating the file etc/translation.vars.custom and re-defining the environment variables there.

Pulling updated translations from Transifex:

1. Pull new and updated translations from Transifex:

   $ tx pull -a
   

2. Correctly format the translations: Do NOT forget this step!

   $ ./update-existing-po.sh
   

3. Look for errors in po files:

   $ ./checkpo.sh
   

4. Find which po files have new strings:

   $ ./findpochanges.sh
   

5. Check in the updated translations:

   # git instructions
   $ git commit -am "Updated translations"
   

6. Check in any new translations: First, look to see which translations are supported in i2pwww/init.py. For any new translations that are NOT in init.py, either delete the po directory i2p2www/translations/xx (if it's not translated enough to add it to the website), or add the language to the table in init.py (if it's translated enough to add it to the website).

   # git instructions
   $ git add i2p2www/translations/* && git commit -am "New translations"
   

Pushing updated translation source (.pot) files to Transifex:

1. Update the .pot files with any changes to the website text:

   $ ./extract-messages.sh
   

2. Check in any changes to the .pot files (optional):

   # git instructions
   $ git commit -am "Updated translation strings"
   

3. Push pots file changes to Transifex:

   $ tx push -s
   

Updating spec tags:

ctags is used to generate references to the specifications. The tags file is i2p2www/spec/spectags. When the specifications are changed, the file should be regenerated and checked in.

Command to generate the file:

```
$ cd i2p2www/spec && ctags -f spectags --langdef=rst --langmap=rst:.rst --regex-rst=/_type-\([a-zA-Z0-9]+\)/\\1/t,type/ --regex-rst=/_struct-\([a-zA-Z0-9]+\)/\\1/s,struct/ --regex-rst=/_msg-\([a-zA-Z]+\)/\\1/m,msg/ -R -n *.rst
```