OpenCFP is a PHP-based conference talk submission system.
Current release: v1.0.3
- Features
- Screenshots
- Contributing
- Requirements
- Installation
- JSON API
- Command-line Utilities
- Testing
- Troubleshooting
- Speaker registration system that gathers contact information.
- Dashboard that allows speakers to submit talk proposals and manage their profile.
- Administrative dashboard for reviewing submitted talks and making selections.
- Command-line utilities for administering the system.
- JSON-API for selected use-cases. (Coming Soon!)
We welcome and love contributions! To facilitate receiving updates to OpenCFP, we encourage you to create a new personal branch after you fork this repository. This branch should be used for content and changes that are specific to your event. However, anything you are willing to push back should be updated in your master branch. This will help keep the master branch generic for future event organizers that choose to use the system. You would then be able to merge master to your private branch and get updates when desired!
- PHP 7.0+
- Apache 2+ with
mod_rewrite
enabled and anAllowOverride all
directive in your<Directory>
block is the recommended web server - Composer requirements are listed in composer.json.
- You may need to install
php7.0-intl
extension for PHP. (php-intl
on CentOS/RHEL-based distributions)
It is recommended for you to always install the latest marked release. Go to https://github.com/opencfp/opencfp/releases
to download it.
Clone this project into your working directory. We recommend always running the master
branch as it was frequent contributions.
Example:
$ git clone [email protected]:opencfp/opencfp.git
Cloning into 'opencfp'...
remote: Counting objects: 4794, done.
remote: Total 4794 (delta 0), reused 0 (delta 0)
Receiving objects: 100% (4794/4794), 1.59 MiB | 10.37 MiB/s, done.
Resolving deltas: 100% (2314/2314), done.
Checking connectivity... done.
OpenCFP can be configured to run in multiple environments. The application environment (CFP_ENV
) must be specified
as an environment variable. If not specified, the default is development
.
An example Apache configuration is provided at /web/htaccess.dist
. Copy this file to /web/.htaccess
or otherwise
configure your web server in the same way and change the CFP_ENV
value to specify a different environment. The
default has been pre-set for development.
SetEnv CFP_ENV production
You will also need to set the CFP_ENV
variable in the shell you are using when doing an install. Here are some
ways to do that with common shells assuming we're using production
:
- bash:
export CFP_ENV=production
- zsh:
export CFP_ENV = production
- fish:
set -x CFP_ENV production
Again, just use your preferred environment in place of production
if required.
From the project directory, run the following command. You may need to download composer.phar
first from http://getcomposer.org
$ script/setup
To run OpenCFP using PHP's built-in web server the following command can be run:
$ script/server
The server uses port 8000
. This is a quick way to get started doing development on OpenCFP itself.
Set up your desired webserver to point to the /web
directory.
Apache 2+ Example:
<VirtualHost *:80>
DocumentRoot /path/to/web
ServerName cfp.conference.com
# Other Directives Here
</VirtualHost>
nginx Example:
server{
server_name cfp.sitename.com;
root /var/www/opencfp/web;
listen 80;
index index.php index.html index.htm;
access_log /var/log/nginx/access.cfp.log;
error_log /var/log/nginx/error.cfp.log;
location / {
try_files $uri $uri/ /index.php?$query_string;
}
location ~ \.php$ {
try_files $uri =404;
fastcgi_param CFP_ENV production;
fastcgi_split_path_info ^(.+\.php)(/.+)$;
fastcgi_pass unix:/var/run/php5-fpm.sock;
fastcgi_read_timeout 150;
fastcgi_index index.php;
fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
include fastcgi_params;
}
}
The application does not currently work properly if you use PHP's built-in server.
Create a new database for the application to use. You will need to have the following handy to continue configuring your installation of OpenCFP:
- Database server hostname
- Database name
- Credentials to an account that can access the above database
Depending on which environment you specified above, you will need to make a copy of the distributed configuration schema to enter your own details into.
For example, if you specified SetEnv CFP_ENV production
:
$ cp config/production.dist.yml config/production.yml
After making a local copy, edit config/production.yml
and specify your own details. Here are some important options
to consider:
Option | Description |
---|---|
application.enddate |
This is the date your call for proposals would end on. |
application.coc_link |
Set this to the link for your conference code of conduct to require speakers to agree to the code of conduct at registration |
secure_ssl |
This should be enabled, if possible. Requires a valid SSL certificate. |
database.* |
This is the database information you collected above. |
mail.* |
This is SMTP configuration for sending mail. The application sends notifications on various system events. |
talk.categories.* |
dbkey: Display Name mapping for your talk categories |
talk.types.* |
dbkey: Display Name mapping for your talk types |
talk.levels.* |
dbkey: Display Name mapping for your talk levels |
For example, if you wanted to setup Mailgun as your email provider, your mail configuration would look something like this:
mail:
host: smtp.mailgun.org
port: 587
username: [email protected]
password: "a1b2c3d4"
encryption: tls
auth_mode: ~
This project uses Phinx to handle migrations. Be sure to copy the phinx.yml.dist
file that is in the
root directory for the project to phinx.yml
and edit it to match your own database settings.
To run migrations, make sure you are in the root directory for the project and run the following:
$ vendor/bin/phinx migrate --environment=production
Note: For updating previously installed instances only run migrations as needed.
- The web server must be able to write to the
/web/uploads
directory in order to - You may need to alter the
memory_limit
of the web server to allow image processing of head-shots. This is largely dictated by the size of the images people upload. Typically 512M works. - Customize templates and
/web/assets/css/site.css
to your heart's content.
OpenCFP has a JSON API (not to be confused with the json-api specification) that can be used by third-party applications to take advantage of a set of features on behalf of a user. The API is enabled by default, but can be disabled if not needed for your instance of OpenCFP.
Configuration for the API is stored under the api
namespace of your configuration YAML file. Currently, there is only
one available configuration setting: whether or not the api is enabled
.
In order to use any of the available APIs in order to do work on a OpenCFP user's behalf, an OAuth2 token must be provided and must have appropriate OAuth2 scope(s) associated with it. Interacting with the authorization endpoints is very much the same as any other OAuth2 implementation; You'll register your custom web application as a Client Application with OpenCFP and from there, you can start to send folks through the Authorization Code Grant Flow in order to eventually obtain a bearer token to act on their behalf.
There are some caveats to the above description that may differ from what you're used to in interacting with the typical OAuth2 implementation:
- Some users you send through the OAuth2 process will not have an account on the target instance of OpenCFP. We take care of that (described below).
- You will not have to create an account on the target OpenCFP process to register your custom web application as an OAuth2 Client Application. We implement a subset of the OAuth 2.0 Dynamic Client Registration Protocol draft to allow applications to dynamically register themselves as Client Applications.
With all of that out of the way, here are some nuts and bolts about our implementation of OAuth2:
- We only support two grant types: Authorization Code & Refresh Token. This allows you to do work on behalf of any OpenCFP user (if authorized) and renew that authorization (bearer token) when it expires.
- Bearer tokens have a time-to-live (TTL) of
3600
seconds (1 hour). Expired tokens will be rejected and you have the option of refreshing or requesting a new token. This may be configurable in the future. - Refresh tokens have a TTL of
604800
seconds (1 week). If you do not want to put users through the authorization code grant flow weekly, have automation rotate you access tokens. - It is highly recommended to only enable this API if you have a valid SSL certificate. OAuth2's security mechanisms are 100% reliant on TLS.
- Authorization endpoints are described below.
This serves as a high-level overview of the OpenCFP API.
Authorization
Authorization endpoints are used as part of the process for obtaining and renewing an Access Token representing a user's authorization for you (as a client developer) to act on their behalf. A step-by-step usage scenario is described for convenience below.
Method | Route | Description |
---|---|---|
GET |
/oauth/authorize |
Starts the authorization flow. |
POST |
/oauth/access_token |
Used to trade an Authorization Code for an Access Token. |
POST |
/oauth/clients |
Client registration endpoint for web application to register as a Client Application. |
Speaker Profile API
The Speaker Profile API allows you to look up information about the currently authenticated user. You might use this to populate attributes in your own custom application based on a user's profile in a target instance of OpenCFP.
Method | Route | Description |
---|---|---|
GET |
/api/me |
Returns JSON body representing information about the authenticated user. |
Talks API
The Talks API allows you to manage the collection of submitted talks for the currently authenticated user.
Method | Route | Description |
---|---|---|
POST |
/api/talks |
Given JSON payload representing a talk, creates talk for authenticated user and issues a 201 Created upon success, appropriate error otherwise. |
GET |
/api/talks |
Returns JSON collection of all talks for authenticated user. |
GET |
/api/talks/{id} |
Returns a particular talk for authenticated user. Returns appropriate responses for unauthorized or non-existent talks. |
PUT |
/api/talks/{id} |
Not Implemented Updates a particular talk. Partial updates are supported through PUT . You are not required to send entire object representation. |
DELETE |
/api/talks/{id} |
Not Implemented Removes a talk. |
In this scenario, we will submit talks on behalf of a user and we make a few assumptions: we assume that you have NOT registered as a Client Application yet and that the user you are submitting talks on behalf of does NOT have an account on the target instance of OpenCFP.
In order to allow third-party clients to register with a target OpenCFP instance as a Client Application, we support a partial implementation of the OAuth 2.0 Dynamic Registration Draft. Specifically, we implement the ability for an arbitrary client to register itself as a Client Application without any pre-arranged authentication process (Software Statements / pre-arranged initial Access Token).
If you haven't previously registered a client application and received a client_id
and client_secret
, you will need to either
do so manually or have your application do so dynamically before being able to redirect users for authorization. Developers using
the client registration endpoint should only register once. It's not going to break anything if you create a new client application
for every single request... but don't do that. It's mean.
To register your application as a Client Application, you will need to send the following request:
POST /oauth/register HTTP/1.1
Host: someopencfp.com
Accept: application/json
Content-Type: application/json
{
"client_name": "Some Custom Web Application"
"redirect_uri": "https://yourwebapp.com/callback"
}
GET https://someopencfp.com/oauth/authorize
Parameters
Name | Type | Description |
---|---|---|
client_id |
string |
Required. The client identifier you received as part of the client application registration process. |
redirect_uri |
string |
The URL in your application where users will be sent after authorizing access. |
scope |
string |
A comma-separated list of scopes. If not provided, scope defaults to an empty list; basically allowing you to authenticate as the user with no authorization to the user's protected resources. |
state |
string |
An unguessable random string used to protect against CSRF attacks. You will send this back when you trade authorization code for access token. |
A user must authenticate to the target instance of OpenCFP before authorizing access to a Client Application. For users that do not have a previously created account, they will have the option of creating a new account. When they complete the account creation process, they will be automatically authenticated into that account and proceed.
[Example Interface Here]
After the user authenticates, they are presented with an authorization interface where they can either approve or deny a Client Application's access to their protected resources. They will see the name of your Client Application in addition to the OAuth2 scopes you have requested. If the user approves, the flow proceeds to the next step. Otherwise, they are still redirected to your application without an Authorization Code and you would need to implement some way of handling that.
[Example Interface Here]
If the user accepts your request, OpenCFP redirects to your site with a code
query parameter as well as the state
you provided. If the states do not match, the process should be aborted.
Parameters
Name | Type | Description |
---|---|---|
client_id |
string |
Required. The client identifier you received as part of the client application registration process. |
client_secret |
string |
Required. The client secret you received as part of the client application registration process. |
code |
string |
Required. The code you received as a response when requesting the authorization code. |
redirect_url |
string |
The URL in your application where users will be sent after authorizing access. |
Response
{
"access_token": "a12834769e4ae7ae178b292c2ee42f710c8316c7",
"refresh_token": "24710c8316c7c2ee42fa1ae7ae178b292834769e",
"token_type": "bearer",
"scope": "public,talks",
"expires_in": 3600
}
Once you have obtained an access token, you can do stuff! You will need to provide that access token for every request you make to restricted API endpoints. All requests SHOULD be sent using TLS (if the OpenCFP instance supports it) because otherwise, we're sending credentials in cleartext.
Access tokens MUST be sent in an Authorization
header as follows from our example access token above:
GET /api/me HTTP/1.1
Host: someopencfp.com
Authorization: Bearer a12834769e4ae7ae178b292c2ee42f710c8316c7
The examples that follow are subject to change. API endpoints and behaviour are still in flux. This should serve more as an example of what to expect as far as interacting with the API, not specifically how the endpoints will work.
View speaker's profile
GET /api/me HTTP/1.1
Host: someopencfp.com
Accept: application/json
Authorization: Bearer a12834769e4ae7ae178b292c2ee42f710c8316c7
HTTP/1.1 200 OK
Content-type: application/json
{
"first_name": "Ham",
"last_name": "Burglar",
"email": "[email protected]
"company": "ACME Corporation",
"twitter": "@hamburglar",
"bio": "..."
}
Submit a talk
POST /api/talks HTTP/1.1
Host: someopencfp.com
Accept: application/json
Content-Type: application/json
Authorization: Bearer a12834769e4ae7ae178b292c2ee42f710c8316c7
{
"title": "Sample Talk",
"description": "...",
"type": "regular",
"level": "mid",
"category": "api"
}
HTTP/1.1 201 Created
Content-type: application/json
{
"id": "1"
"title": "Sample Talk",
"description": "...",
"type": "regular",
"level": "mid",
"category": "api"
}
Verify talk was submitted
GET /api/talks/1 HTTP/1.1
Host: someopencfp.com
Accept: application/json
Authorization: Bearer a12834769e4ae7ae178b292c2ee42f710c8316c7
HTTP/1.1 200 OK
Content-type: application/json
{
"id": "1"
"title": "Sample Talk",
"description": "...",
"type": "regular",
"level": "mid",
"category": "api"
}
Delete talk
DELETE /api/talks/1 HTTP/1.1
Host: someopencfp.com
Accept: application/json
Authorization: Bearer a12834769e4ae7ae178b292c2ee42f710c8316c7
HTTP/1.1 200 OK
Access tokens have a TTL of one hour. Once expired, you will need to either request another access token or we give you the ability to refresh access tokens through use of the Refresh Token Grant. Taking advantage of this is actually pretty simple. You'll basically send a request that looks similar to this:
POST /oauth/access_token HTTP/1.1
Host: someopencfp.com
Accept: application/json
Authorization: Bearer a12834769e4ae7ae178b292c2ee42f710c8316c7
grant_type=refresh_token&refresh_token=24710c8316c7c2ee42fa1ae7ae178b292834769e
HTTP/1.1 200 OK
Content-type: application/json
{
"access_token": "2ee42f710c8316c7a12834769e4ae7ae178b292c",
"refresh_token": "1ae7ae178b292834769e24710c8316c7c2ee42fa",
"token_type": "bearer",
"scope": "public,talks",
"expires_in": 3600
}
After refreshing the access token, you'll obviously want to update the previous token you've associated with your user. Also note that refresh tokens are rotated in addition to the access token. You'll want to keep track of this per-user.
OpenCFP comes bundled with a few command-line utilities to administer the system. A full list of commands (along with help for each) can be found by running the following in the project root:
$ bin/opencfp
Administrators are authorized to review speaker information in addition to specifying talk favorites and making selections.
Adding [email protected]
to the admin group:
$ bin/opencfp admin:promote --env=production [email protected]
Removing [email protected]
from the admin group:
$ bin/opencfp admin:demote --env=production [email protected]
Users are needed for you system, and sometimes you want to add users via command line.
Adding a speaker:
$ bin/opencfp user:create --first_name="Speaker" --last_name="Name" --email="[email protected]" --password="somePassw0rd!"
Add an admin:
$ bin/opencfp user:create --first_name="Admin" --last_name="Name" --email="[email protected]" --password="somePassw0rd!" --admin
OpenCFP uses Twig as a templating engine and HTML Purifier for input filtering. Both of these packages maintain a cache, if enabled. If you need to clear all application caches:
$ bin/opencfp cache:clear
OpenCFP follows the Scripts to Rule Them All pattern. This allows for an easy to follow convention for common tasks when developing applications.
This command will install all dependencies, run database migrations, and alert you of any missing configs.
$ script/setup
This command will update all dependencies and run new migrations
$ script/update
This command will start a built-in php web server, using port 8000
.
$ script/server
This command will run the PHPUnit test suite using distributed phpunit config, phpunit.xml.dist
, if
no phpunit.xml is found in the root.
$ script/test
There is a test suite that uses PHPUnit in the /tests directory. To set up your environment for testing:
- Create a testing database, and update the name and credentials in /config/testing.yml
- Copy the default
phinx.yml.dist
tophinx.yml
- The recommended way to run the tests is:
$ script/test
The default phpunit.xml.dist file is in the root directory for the project.
I'm getting weird permissions-related errors to do with HTML Purifier.
You may need to edit directory permissions for some vendor packages such as HTML Purifier. Check the /cache
directory's
permissions first (if you have cache.enabled
set to true
).