User:Mihaineacsu/proposal2014

Student Information

Name: Mihai Neacșu
Email: mihai.mneacsu@gmail.com
Telephone: +40 755 283 297
Time zone: UTC+02:00
IRC:  mihaineacsu
Source control username on github: mihaineacsu
Source control username on sourceforge: mihaineacsu

University Information

University: Politehnica University of Bucharest
Major: Computer Science
Current Year and Expected Graduation date: 4th year, 2014
Degree: BSc

Abstract

BRL-CADs material database will help users list, share, download and load material information and have that information be seamlessly integrated with BRL-CADs modelling (Archer) and analysis (gqa) tools. Part of this project will consist of creating a centralised repository where users can perform the above mentioned actions. It will provide users the means for searching, sorting/filtering, specifying, storing, importing, exporting, and reusing material properties. The second part of the project will consist of creating a command-line tool that uses the online material database/repository in order to get/put material data, create .density files (and other formats), or add material data to a .g geometry file. Further on, emphasis will be put on integrating it with BRL-CAD such as making it a part of the Geometry Editing Library (libged).

Detailed description

Server-side

Will provide a REST(Representational state transfer) API. Why REST? Because REST is a simple way to organize interactions between independent systems. It allows us to interact with minimal overhead with clients.

How does it work? We operate on URLs and each URL identifies a resource. These are exactly the URLs which are assigned to web pages.

So for e.g.. performing a GET request on mater.brl-cad.org/list will provide us a with a JSON that the client will interpret and show to the user in an organised way. The client can either be a running BRL-CAD instance or the browser.

{

 "type": “BRL-CAD Materials List",
 "format-version": "1.0.0",
 "materials": [
   { <material-details> },
   { <material-details> }
 ]

}

where material-details JSON is: {

 "name": “Iron",
 “id”: 3432,
 “traits”: [
   “Density”: “7.87 g/cm^3”,
   “Specific Heat”: {“value”: “0.12 cal/g C”, “temperature”: 100 C”}
 ]
 "short": “Short description",
 "screenshot-url": "",
 "version": "1.0",
 "authors": [
   {<author-details>},
   {<author-details>}
 ]

}

where author-details JSON is: {

 "author-name": "Weiss Richard",
 "author-contact": "some free text, probably email",
 "author-url": "http://a.nice.site/with/a/homepage.html"

}

This is quite bloated but a simple version listing just the materials and their properties might be more suitable when accessing the REST API from BRL-CAD.

Proposed REST API:

GET mater.brlcad.org/list - list all available materials
GET mater.brlcad.org/list?id=MATERIAL_ID&name=MATERIAL_NAME - search and list by MATERIAL_ID or MATERIAL_NAME
POST mater.brlcad.org/remove?id=MATERIAL_ID - remove MATERIAL_ID, only available to some privileged authenticated users
POST mater.brlcad.org/update?id=MATERIAL_ID - add or update MATERIAL_ID, only available to some privileged authenticated users

The last 2 resources require POST actions because they need credential checks in the process. The ability to update/remove a certain material would only be available to some groups of users such as admins or the author.

We’ll need to setup an an admin interface in order for some privileged users to easily review, add, update, remove materials.

I’ve noticed existing work on material database is using XML format to display available materials, I would suggest using JSON as it is a bit lighter, human-readable, but that’s more of a personal preference and just an implementation detail. Exporting the data can be done in either format, but we might want to consider the lightest option for BRL-CAD parsing (there are number of XML/CSV/JSON parser for C++ out there but I haven’t tested any of them performance wise yes - I don’t really know if it could a be strain of performance if there is a large number of materials and we reach a point where the user can browse available options in BRL-CAD).

Client-side

Clients will initially request a list of materials from the Server. They will interpret the JSON and display to include informations such as Author, Description, and most importantly their traits.

The materials list will be accessible from browser - a dinamically generated page will list all the materials and offer options for authenticate in order to add new materials. BRL-CAD - initially just a command line tool that talks to the website to get/put material data, create .density files (and other formats), or add material data to a .g geometry file

HTTP requests on BRL-CAD can be made using either libcurl to retrive data and jsoncpp to parse the JSONs if we to decide to opt for a JSON format. I would look into what’s fastest/most stable out there.

Timeline

June
Week 1: Decide details on server implementation and start working on providing a REST API
Week 2: Have a fully working API implemented and create a test suite for it
Week 3: Create an admin area for managing the available resources
Week 4: Set up website that lists and enable users to perform actions such as searching, viewing, adding materials


 July Week 1: Start working on BRL-CAD command-line command
Week 2: create .density files (and other formats),
Week 3: add material data to a .g geometry file
Week 4: other API routines and integrate in libged

August
Week 1: Continue working on previous point
Week 2: Have a fully working interface with a passing test suite
Week 3: Refactoring & documentation
Week 4: Refactoring

This provides a rough guideline of how the project will be done. As you will see, I plan to start working right after the accepted proposals list is published. Because of that, according to my timeline, I will have finished the project earlier. I don't know if this will be the case, but in all events, I will have time to cover any unexpected issue that may come up.

Bio

I’m Mihai, a CS undergraduate at Politehnica Univ. of Bucharest, currently in the 4th and final year of study.

I believe each of us needs to leave his mark in some way, for me it has been about making software that is used by as much as many other people. I didn't have any real opportunity to do so, until I found out about the Open-Source Development Course, taking place in our university. I applied and after a series of interviews I was accepted. It's a course that helps students make their first contributions within an open-source project and get a feeling of what real world software development means. I got to learn about working in an open-source community, using version-control systems like git, unit-testing, asking for help on mailing lists and other things like that.

One of the open source projects I got to contribute to was Jeopy, a local-network multiplayer quiz game written in Python with PyQt (Python binding for Qt) that features trivia in computer science and was used in some of the first year's classes for the undergraduate students at my university in order to help them recap course material. Four years later, I'm still participating in the Open-Source Development Course but as a mentor.

I’m actively involved in university life, I am an undergraduate teaching assistant for a few classes and I try to participates in technical paper presentations as often as I can.

I am the first one to sign up for hackathons as I strongly believe them to be a unique chance to get together with people who also share my passion in building something new and want to learn from each other in the process.

Last summer I was an an intern at the romanian start-up uberVU. I continued to work there for another 7 months after the internship was completed. uberVU is a social media analytics applications and my work was mainly focused on backend services (mostly Python), where I had to implement a few systems and ensure they’re scalable and robust.

I volunteer a few hours week at Hospice of Hope, a palliative care center. I also volunteer for activities organised by the Romanian Open Source Education group which I’m also a part of.

I’m a disciplined individual and I try to organise myself as best as I can though that doesn’t always happen as I always get involved in a big number of projects. That would not be the case this summer as I’m finishing university and considering tackling just a couple of open-source project, BRL-CAD being one of them.

Why did I choose to apply for BRL-CAD?

To be honest, I’m not an avid user of CAD software. I haven’t used one since high-school, at some time back then I did get into 3D modelling and photo editing and many many other activities because I was pretty much clueless of what I wanted to do so I experimented with every kind of piece of software I could get my hands on. I got to work with AutoCAD and Cinema 4D and it was quite insightful stuff on how 3D modelling is done. I’m particularly fond of this project because it involves working on the web platform and on BRL-CAD and I’d love to work on tying them together. I feel that getting involved in a project of this magnitude will help me bring my skills to a higher standard. The community was open, friendly and helpful and that’s also a big plus. In a nutshell, this project is, to me, the perfect opportunity to learn new things, to continue my development as a programmer and to do what I've always wanted: contribute with a piece of software that would actually help the community.