Skip to main content

How I did customize "rasa-nlu-trainer" as my own tool


Check out my implementation here

Background

I wanted to have a tool for human beings to classify intents and extract entities of texts which were obtained from a raw dataset such as Rocket.chat's conversation, Maluuba Frames or here. Then, the output (labeled texts) could be consumed by an NLU tool such as Rasa NLU.

rasa-nlu-trainer was a potential one which I didn't need to build an app from scratch. However, I needed to add more of my own features to fulfill my needs. They were:

1. Loading/displaying raw texts stored by a database such as MongoDB
2. Manually labeling intents and entities for the loaded texts
3. Persisting labeled texts into the database

I firstly did look up what rasa-nlu-trainer's technologies were used in order to see how to implement my mentioned features.

At first glance

rasa-nlu-trainer was bootstrapped with Create React App. Create React App is a tool to create a React app with no build configuration, as it said. This tool is also recommended by the official React.js tutorial. I actually was very new to React.js. I wondered what problem Creat React App tried to solve so that I first wanted to set up a project without Create React App. I followed this post.

Well, let me say if I built a React app from scratch then I needed to do some manual configurations for Webpack and Babel including webpack.config.json and .babelrc

Then, I tried to use Create React App, and I followed this guide. All right, let me say Create React App helped me create a React boilerplate project with hiding Webpack and Babel's configuration so that as a developer I only needed to focus on writing production code.

The nuts and bolts

I went on starting to watch how rasa-nlu-trainer's source code was organized and what its techniques are used.

Folder structure

1. Auto-created stuff by Create React App
   - public/index.html
   - src/index.js
2. Overrides
   - They didn't use "react-script", they created their own way instead! That was the reason why some additional folder/files were added such as server.js, scripts and config.

Used technologies

- Programming language: Javascript
- Babel: ES6 transpiler
- Project bundle: Webpack
- Dependencies management: npm
- Backend: Node.js
- Fronted framework/project-based: React.js
- React model storing: React-Redux
- React UI extension components: Ant Design
- TypeScript

Hooray! Welcome me to Javascript world!

Architecture

- index.js: Provider{store} <---> LocaleProvider ---> App
- How "separation of concerns" works: import and  export
- Global data (store) is immutable and only changed by files actions.js and reducer.js

Proposed changes

Finally, I could start to add my needs to the existing rasa-nlu-trainer app. I listed out what my features:
- Creating an overview page as a custom React component to display all loaded texts
- Customizing the existing page for classifying intents and extracting entities.
- Implementing Node.js methods to load/stores texts (as JSON)  which stored in MongoDB

The rule of thumb was I needed to follow the mentioned architecture for modifying the global data which were managed by React-Redux.

You can take a look at the changes from this pull request.

Conclusion

Sometimes, the best way to learn new techs is to get down into an existing tool and customize it with your own needs. 💪

Comments

Popular posts from this blog

Safari (older than 14.1.1) rejects TLS connections of TURN with Let's Encrypt certificates

Summary We could not make a video call successfully on Safari older than 14.1.1. The call kept disconnecting for some seconds. We got the following error in Nginx’s log SSL: error:14094418:SSL routines:ssl3_read_bytes:tlsv1 alert unknown ca:SSL alert number 48   Root cause The Safari didn’t trust our TLS certificates (certified by Let’s Encrypt) when establishing a TLS connection into our TURN server. It is a known issue that could be found at Bug 219274 - ICE does not resolve for `turns` relay candidates rooted in LetsEncrypt CA . The issue was just recently fixed (end of 2020), therefore old versions of Safari still meet the issues.   Solution In our development environment, we replaced the untrusted certificate with a trusted one. We will warn our customers about this issue in our installation guide. Technical explained WebKit is a browser engine developed by Apple and primarily used in its Safari web browser, as well as all iOS web browsers. The WebRTC of WebKit relied on ...

DevOps Toolchain Enhancement

 Historically, our company ubitec had started with a customer project. Agile/Scrum was our proposal for working with customers. Time by time, Agile/Scrum also became our culture for software development. To be successful with this development approach, we somehow needed to have a fast release for customers (i.e. every one week). Back then, we had a build tool Jenkins which was responsible for having sprint release packages for our customers. The build job pipelines contain some steps such as gathering the artifacts, checking the code convention, running the tests, building docker images, and packaging an archived file (a zip file). The set of tools involved in a pipeline is roughly called a toolchain. It is just a part of a bigger process called the DevOps toolchain. Source: https://www.ibm.com/blogs/cloud-archive/2016/11/devops-architecture-available-on-bluemix-garage-method-site/ DevOps is a proven method that fits Agile. Today,  it is even treated as a mandatory factor...

Why Agile and Scrum Matter

Changes is constant, software development as well. Do you know the development of mobile phones, starting from Motorola DynaTAC (1973) to Iphone 6 (2014). How about the software? The new versions are released with new improvements. Software is not always able to automatic repairing. Therefore, software should be changed frequently and we need responding to changes. source:  http://answers.microsoft.com/ In working software today, how about customers' collaboration or requirements? We ignore the fact that many customers don't know what they want . We ignore the fact that when they know what they want, they can't describe it . We ignore the fact the even when they can describe it, they often a proposed solution rather than a real need . We ignore, that a lot of customers g ive us a solution but not the problem . source:  http://accelerateddevelopment.blogspot.com/ And, in working process, earlier founded bugs are cheaper. Manifesto for Agile ...

Sharing a virtualenv across several Python projects using Pipenv

There is a standard library for all projects in Python. However, several projects don’t always have the same dependencies all the time. That is where virtual environments come to play. You can follow this official document to use two separated tools  virtualenv and pip to  fulfill that need. My preferred alternative is to use pipenv . Pipenv is easy to use and convenient. The following are my steps to make a shared virtualenv for my all projects which requires the same dependencies. Step 1. Create an isolated virtualenv. python -m venv my-shared-env Step 2. Create a symbolic link to the created virtualenv. cd project_1 ln -s ~/.local/share/virtualenvs/my-shared-env .venv I have encountered the following issue at step 1. FileNotFoundError: [Errno 2] No such file or directory: '{my_project_path}/.venv/bin/pip': '{my_project_path}/.venv/bin/pip' The root cause was I tried to create virtualenv by running pipenv install and renaming the generated virtualenv to ...