[DEBUG][19:45:47] Files that will be checked:
[DEBUG][19:45:47] coala is run only on changed files, bears' log messages from previous runs may not appear. You may use the `--flush-cache` flag to see them.
Executing section generalization...
[WARNING][19:45:48] Failed to read file '/home/travis/build/coala/community/.ci/deploy_key.enc'. It seems to contain non-unicode characters. Leaving it out.
[WARNING][19:45:48] Failed to read file '/home/travis/build/coala/community/activity/__pycache__/scraper.cpython-36.pyc'. It seems to contain non-unicode characters. Leaving it out.

They need to be excluded. Probably others also; see .gitignore.

As the contents of the page include a static snapshot of org activity , it is important to know when it was last generated in UTC, and show how long ago it was generated , aware that the readers timezone is likely not UTC.

https://coala.io/#/getinvolved is part of https://github.com/coala/landing-frontend , and it uses data collated from the webservice provided by https://gitlab.com/coala/landing .

The same backend data should be imported into a data model in this repo, with a similar UI, but UI is not so important.

Note: The data from the webservice is a bit out of date. Not a problem for this task.

After #59 , we need to import gsoc data from 2016.

This is more complicated, as coala was a sub-org of PSF. Doing this generically could be quite difficult.

We could search for projects prefixed with [org name]: .
There are other orgs which used similar naming conventions, like Astropy: ... within OpenAstronomy

See https://www.openhub.net/orgs/coala/ , and the OpenHub apis. We need an OpenHub app that collects all of the data they make available for use by this website.

Netlify forms can be used to let people 'become a developer' from their web browser. It allows signing using GitHub, for authenticated collection of their username to be added to the org. We could also add a field for the issue they want to be assigned to.

It is then easy for maintainers to invite them to the org after some rudimentary checks, or add a http://webtask.io which automatically invites them if they pass some basic automated checks to prevent abuse.

Probably a good app name is '/join'

Using IGitt's write_repositories allows easy access to the repositories, but it is not necessary.
This may need upstream enhancements.

Related to coala/coala-html#143 (comment)

Personally, I think it's preferable to use Travis's GitHub deployment feature rather than relying on a pretty hacky script.

https://github.com/coala/community/blob/master/docs/_static/images/header.png needs to be removed.

This image was added as a temporary solution for #10 , as it was a newcomer issue.

This project is a re-usable component that can help any community, so the README should not focus on coala.

It should be replaced with a project specific logo which symbolises creating/supporting/growing communities.

I think want at least the coala maintainers, coala developers and coala newcomer teams to have a page, with a list of team members similar to https://coala.io/#/getinvolved .

I think it is OK to have a page about all teams in https://github.com/orgs/coala/teams .

However we also need a way to name our temporary teams so that they do not appear on the website.

e.g. @adtac created a team to help him manage inactive newcomers at coala/meta#74 , so they could be removed. Maybe we dont want a team like that appearing on the website ;-)

1:13:24 AM: /usr/local/bin/run-build-functions.sh: line 150: /opt/buildhome/python3.6.2/bin/activate: No such file or directory
1:13:24 AM: Error setting python version from runtime.txt
1:13:24 AM: Will use default version (2.7)

https://app.netlify.com/sites/coala-community/deploys/5aa17d3e92226a6064ec5772

Since the Coala Community site is Open Source, I think it'd be a good idea to encourage contributions to it by adding a 'Fork Me On GitHub' ribbon to each page!

The ribbon, which could look something close to these would take the viewer to https://github.com/coala/community/fork where they'll be prompted to fork the repository and make changes :) I think it's important to have this as the site is about the community around Coala and GCI, both of which are founded around Open Source contributions 😄

Examples of ribbons/links like this are often found on stuff like the Docsify Site and the Jekyll Documentation.

As this repository should be usable for any org, detect the name of the GitHub org from the Travis environment variables, and fetch the org's logo.

And then convert the logo into a favicon.

When webservices will be up it will provide all the mrs of coala on GitHub and GitLab through an API,
that should be imported here to build the gamification app.

As the website has to cater with unexpected data , the website build scripts will be permissive, logging errors rather than failing.
Currently the Travis build logs are the best spot for logging.

http://community.coala.io/static/tasks.yaml contains all tasks, sorted by task identifier.

Create an RSS feed of tasks, sorted by last_modified

The tasks in the GCI system has links to issues, and the GCI app uses the assignee to link GCI students with the issues they are working on , or have completed.

However many of our sub-orgs do not invite GCI students into their org, and on GitHub that means they can not be an assignee. They need to be detected using the pull requests.

See https://travis-ci.org/coala/community/builds/317909190#L772

I broke the logo (#1) when moving the GCI stuff to /gci.
I forgot to move the logo from /gci to /.

The twitter feed for the organisation should be fetched and displayed on the Google Code-In community website within a Twitter timeline (obtainable from https://publish.twitter.com/). The Twitter handle for the organisation should be fetched via the Google Code-In api (see https://codein.withgoogle.com/api/program/2017/organization/) and used in conjunction with the embedded timeline obtained from Twitter in order to show tweets from the correct Twitter account. If the Twitter handle for the organisation is defined as NULL within the Google Code-In api response, the website should fail gracefully and not display the Twitter timeline.

When a task has max_instances: 1 and an issues URL , it can be annotated with an extra attribute to indicate the task status: available, claimed or completed.

Instance are exported as instances.yaml , and indicate whether the task is in progress or completed.

Completed issues are the easiest to work with.

In progress tasks are slightly more complicated because the publish instances are filtered to remove tasks still in early phases of the task lifecycle. See https://github.com/coala/community/blob/master/gci/students.py#L84

GCI has an API.

There is also a 'program' API for GSOC information at https://summerofcode.withgoogle.com/api/program/ ; the same data is available at https://summerofcode.withgoogle.com/api/program/2017/

That only provides information about the completed 2017 program, which is worth including, as preparation for the 2018 program.

For the 2018 program, there is an ical which can also be imported into the gsoc data model so we have at least basic information about the next GSOC program able to be used in our community website.

A proper documentation with installation guide could help a lot to setup locally and contribute.

The .coafile should check the file types already in this repository,

and it should reject the word 'coala' appearing in any file.

The readme needs to say:

• This repository is a static django website
• It uses the GCI private API
• But it only re-publishes publicly available information
• It uses IGitt to support both GitHub & GitLab

For GCI task which did not have a PR assessment item, the most recent attachment on approved GCI task instances are likely to be the outputs of the task, and should be fetched into the private folder.

The fetcher should not push these private artifacts to the public folder to be published. This needs to be manually done to ensure those artifacts had correct licensing and were indeed the correct output of the task.

A naming convention needs to be established so that those attachments moved to the public folder by an org admin will be linked to the task, and probably presented as a gallery.

A coala-specific implementation detail, maybe best dealt with as a separate enhancement, is some of the coala tasks required uploading the artifacts to Behance or DeviantArt, and those links need to be extracted from the task instance instead of attachments.

There have been a few failures to build, which havent been seen because python manage.py distill-local doesnt return a non-zero exit status.

#19 is one such occurrence, which was with Django 2.

Here is the same occurring when using Django <2.0

https://travis-ci.org/jayvdb/coala_community/builds/311125722#L1182

We're facing a lot of issues right now because of config files. Creating a config file is required to maintain details like org_name, twitter handle, website link etc that could be used further.

We can make use of the generic ListView to provide a default viewer of datasets that we are importing.

https://developer.mozilla.org/en-US/docs/Learn/Server-side/Django/Generic_views

In 2017, coala was gsoc org 5817061024464896.

As this repository is generic, the tool must find the project identifier using only the org_name which is exposed in the community app.

Scrape the list of org projects from that page , e.g. 5154725527814144, pull in relevant data, such as student and mentor display name.

The scraper must be part of the django system so that any GSOC org can load their 2017 data using the scraper.

https://github.com/publiclab/community-toolbox has a similar intention as this repo + https://github.com/coala/gh-board .

We should extract feature ideas and put them into the relevant coala project.

Currently in https://github.com/coala/community/blob/master/gci/config.py#L6 we define GCI_DATA_DIRECTORY locally using python file management system. Since there is global setting to use this specified here. So we want to replace that to use the global setting.

label difficulty/low newcomer

Update .gitignore for ignoring the .idea dir and .DS_Store

Description
A way to get participants’ geolocation is important for any org to show its global influence, this can be used for map display and statistics.

Purpose
To show how people from different regions can together join open source projects and make the project great together so more people might be encouraged. Moreover, it can strengthen the local bond in open source project development. It can also serve for statistical purposes.

Implementation
To acquire all geolocation of participants, we can use the following procedure:

Names(from google) -> Github account -> Location(curl -i https://api.github.com/users/grox-ni/location) -> Recognization(with some external API to recognize geolocation from string).->Display(e.g. D3.js)

Anyone who wants to can disclose their approx geolocation on GitHub as a string, and it can be converted to geo coords.

.coafile currently relies on explicit inclusion of new apps, which means by default new code isnt checked.

.coafile doesnt invoke the pyflakes checker.
A few problems will need to be fixed as part of this change.

Website may include a bar chart widget, which will show the dynamics of solving & creating issues. It will highlight how active the community has been during periods of time.
The X-axis of the chart will be timeline (e.g. last 5 month), and Y-axis - amount of issues solved and created. To be visually intuitive, created & solved issues should be marked with different colors. Also, dates with positive tendency (rise in activity) may be marked green, dates with negative tendency - red.
Tools:
Bar chart creation: http://www.chartjs.org or https://d3js.org
Fetching data: IGitt has issue searching func. For each repo of the org, we can search issues by last update date and status (it must be closed).

We can import data from Gitter Developer API which can be useful in creating metrics for Newcomer metrics and gamification project.

.travis.yml currently sets COMMIT_AUTHOR_EMAIL: "[email protected]" , but that violates this repositories rule that 'coala' shouldnt be mentioned anywhere.

One approach is to use a generic email address.
Another is to use the git remote info to find the GitHub org (#26), and then derive an email address from the GitHub org.

Split off from #1

Using travis environment variables will work in Travis, but not locally, or on other build. bad idea.

We should use git remote -v to fetch the remote name (there should be some IGitt functionality for this?), and then use https://github.com/retr0h/git-url-parse to extract the org name.

This functionality belongs in the main application, not in the gci app.

See http://lab.lepture.com/github-cards/ for how to easily create a GitHub Card.
As the list of users is dynamic, the solution needs to be dynamic.

We can export a subset of the data as yaml/json, and store it in the repository, so that people can develop without access to a GCI API token

The student profile section isn't rendering the HTML code.

Instead of print statements to the travis log, problems should be logged properly and a view generated with the logs.

Upstream meeb/django-distill#7

When webservices will be up it will provide all the issues of coala on GitHub and GitLab through an API,
that should be imported here to build the gamification app.