Adventure archive is a planning and research tool for the amateur adventurer and traveller. It is designed to be a complete archive of the old Lonely Planet Thorntree travel forum, a goldmine for niche travel information on far away places. It features an easy to use API and front-end interface for researchers.
Acquire Adventure Archive
Adventure archive can be cloned from the public GitHub repository. This repository is also mirrored on GitLab where a complete continuous integration and deployment pipeline can be found. This pipeline automatically runs an extensive test suite, ensuring the application is robust and bug free on every production commit. The results and coverage of these tests can be viewed on the coverage report page.
After robustly testing and linting the source code, the CI/CD pipeline automatically minifies and compiles the entire front-end into a single production-mode bundle file, as well as packaging the application back-end in a docker container. This container is then published via a container registry to a heroku webserver where it is then deployed publicly. More information can be found below.
Installation
Once the source code is acquired, Adventure Archive can be easily installed. Node and NPM are the only big-picture dependencies, and must be installed on your system. The remaining JS dependencies can be installed with the usual npm install. The front-end bundle can be compiled using npm run build:frontend. Please note that by default this will compile in development mode, as the webpack mode flag has been set to development in the source. (As part of the live deployment, the CI/CD pipeline automatically switches this to production for a minified bundle.) If you would like to automatically build for production, you can instead use npm run build:production. Running npm run watch will automatically track and compile changes in the front-end code.
The webserver can be started with npm start, and can continually track changes using npm run dev script. Both the front-end and back-end code can be tracked, recompiled, and restarted automatically by the aggregated script npm run dev:full.
The CD/CI pipeline can be investigated locally using the npm run test and npm run lint commands. Additionally a complete JSDoc's documentation can be compiled by running npm run doc. This documentation can be viewed in documentation/index.html.
Finally, should you wish to extend the database's collection and continue the project, you should investigate the npm run scrapeData, npm run speedTest, and npm run rootIndex commands. These scripts are interesting, and their source code shows how the database was built (Relevant to the project, so I decided to leave them in my submission for your viewing), but are not run in the production container.
Deployment
Adventure Archive can be found deployed live at adventure-archive.herokuapp.com. In the navigation bar, you will find a search bar where you can search comments by their contents, as well as a drop-down advanced search panel. This advanced search panel allows you to specify multiple text queries (AND), multiple countries (OR), multiple sub-forums (OR), and a maximum number of comments to query for. Please note that some obscure searches may take a number of seconds to query, and as such please observe the small loading indicator next to the search bar; This indicator will remain spinning for as long as work is being done behind the scenes (NoSQL database is being queried).
Once the results of your search have been returned, you may preview them at will. If you find one which interests you, you may click on the forum title ("First-Hand Reports from Japan" in the above example), to view a full-screen view of the entire forum which the comment was posted in. Should you be interested in other similar posts you can click on the country, sub-forum, and user headings in order to quickly search for more comments with similar attributes.
Should you wish to add first hand information of your own to the repository of knowledge or simply correct misleading or out-dated information, you can easily post your own comment to a forum. Simply "login" with a username of choice by clicking the LOGIN button on the top left of the navigation header, and enter a username to use. Then simply open a forum up, scroll to the bottom and post your comment.
API
Should the front-end of Adventure Archive not be to your taste, you can build your application to utilise Adventure Archive’s public API. All endpoints listed below are extensions of the root URL https://adventure-archive.herokuapp.com
Available versions
/api/v0.0.1 The latest and greatest API version.
/api A default API route which will automatically route to v0.0.1.
All endpoints listed below are extensions of these API endpoints.
Available Resources
GET /comments Will return a list a list of comment documents. These comments will contain a unique _id, as well as their content, date posted, the threadId of the thread they were posted to, and the userId of the user who posted them.
Comments can be filtered in a number of ways. These query parameters can be combined together in as many combinations as you chose, however detailed specific queries will perform much slower than simpler queries.
/comments?thread=[ID] : Comments belonging to a specific thread.
/comments?search=[SEARCH QUERY] : Comments containing a specific sub-string.
/comments?user=[ID] : Comments posted by a specific user.
/comments?country=[COUNTRY NAME] : Comments about a specific country
/comments?subforum=[FORUM NAME] : Comments posted to a specific sub-forum
/comments?groupByThread=true : A maximum of one comment per thread
/comments?limit=[NUMBER] : A maximum of [NUMBER] results will be returned
/comments?random=[NUMBER] : Simmilar to limit, but comments will be better randomly sampled.
The above queries can be combined using &, so to search for up to 10 comments posted to threads about Canada OR Mexico, containing the words Vancouver AND Hiking the following endpoint would be appropriate:
/comments?country=canada
Please note that all searches are case sensitive and substring based.
A specific comment can be accessed at the end point
/comments/[ID]
New comments may be posted to the endpoint
POST /comments
Post bodies must contain an ASCII string comment with a minimum length of 1, a valid unique ID of a user, and a valid unique ID of a thread.
GET /users Will return a list a list of user documents. These comments will contain a unique _id, as well as their name. The following endpoints are available for users, with similar filtering rules as above.
/users?name=[USER NAME]: Search by name
/users?limit=[NUMBER] : Limit query results
/user/[ID] : Get a specific user document
New users may be posted to the endpoint
POST /users
Post bodies must contain an ASCII string name with a minimum length of 1. A unique ID will be created based on a sha256 hash of this name, so all name’s must also by default be unique. Should a user already exist with this name, a record of that user will be returned.
GET /threads Will return a list a list of thread documents. These comments will contain a unique _id, as well as their title, country, and subforum. The following endpoints are available for threads, with similar filtering rules as above.
/threads?title=[THREAD TITLE]: Search by title
/threads?country=[COUNTRY NAME] : Threads about a specific country
/threads?subforum=[FORUM NAME] : Threads posted to a specific sub-forum
/threads?limit=[NUMBER] : Limit query results
/threads/[ID] : Get a specific thread document
All available resources additionally have a links attribute containing HATEOAS style links for better discovery. Sensible validation is applied to query parameters to prevent attacks, but in general, the API is very forgiving, returning what it can, when it can.
Citations and Sources
The Adventure Archive favicon is an adapted version of an ICON from Font Awesome
The content of this archive has been archived from the Thorntree forum. This information was retrieved from the compact form api endpoint which can be found by appending forum URL's with ?compact. An example of which can be found here