https://github.com/Nithin-Kamineni/Shellhacks-backend

Inspiration - To implement an efficient natural language model(spaCy) to extract attributes of the description scene in the data provided and show the mapping between these timestamp and attributes identified in a graph visually on the frontend.

What it does - The model named spacCy extracts the unique attributes in the scene. Later a csv file is created mapping each timestamp with a quantified field of each of the unique attributes identifies. Then an API call written in flask sends this mapping between each of the timestamp and unique attributes to the frontend.

How we built it -

We have used the spaCy library to analyze a given text describing a city street scene and extract relevant entities while mapping them to their associated quantities. It begins by loading the pre-trained "en_core_web_sm" model for English language processing. The text is then segmented into sentences, and a new spaCy model is initialized to add custom entity recognition rules. The entity ruler component is used to define patterns for recognizing entities, including objects like "person," "car," "building," and quantities like "one," "several," and "none." Training data is prepared by tokenizing sentences and extracting recognized entities, which are later filtered and used to create a mapping of objects to their quantities. This mapping helps understand the quantitative aspects of the described scene, such as the number of cars, people, or buildings. Later we write a script to generate a csv file showing the mapping between each timestamp and a quantified field of each of the unique attributes identified. Then we used flask on the backend to write API calls which send the mapping between each of the timestamp and unique attributes to the frontend.

Challenges we ran into - To come up with right NLP algorithm whcih analyzes the data corretly to help us find the

                                         correct  attributes from the scene
                                       - To increase the efficiency of the NLP algorithm.

Accomplishments that we're proud of - We were able to get the attributes which have a predominant influence in the

                                                              given scene descriptions
                                                             - We were able to use an efficient NLP algorithm(spaCy)

What we learned - We learned building API using flask, We were able to analyze how to figure out the efficient NLP algorithm to use.

What's next for https://github.com/Nithin-Kamineni/Shellhacks-backend

We can further explore on the NLP models available and come up with the most efficient one.

Built With

• flask
• natural-language-processing
• postgresql
• python
• sql