This commit adds the `searchwords_for_filtering.py` file, which contains a list of search words related to electric vehicles. These search words will be used for filtering and categorizing content related to electric vehicles.
This commit refactors the ArangoDB class and script for fetching electric vehicle (EV) speeches. The changes improve the efficiency and readability of the code, ensuring a smoother retrieval process for EV speeches.
feat: Add Ollama sentiment analysis and argument generation to speech processing
This commit adds functionality to the `llm_sentiment.py` script for sentiment analysis and argument generation using the Ollama library. The script now generates sentiment scores for each speech and stores them in the `llm_sentiment` field of the speech dictionary. Additionally, the script generates a list of arguments related to electric vehicles mentioned in each speech and stores them in the `llm_arguments` field. Finally, a detailed summary of what is said about electric vehicles in each speech is generated and stored in the `llm_summary` field.
These changes enhance the speech processing capabilities by providing sentiment analysis and extracting relevant arguments and summaries related to electric vehicles.
This commit updates the .gitignore file to include all Python files by using the `*.py` pattern. This change ensures that all Python files in the repository are tracked by Git.
Refactor the ArangoDB class and the script for fetching documents related to electric cars. The code has been modified to improve readability and maintainability. The script now uses the Ollama library to generate arguments based on the fetched documents. This update will make it easier to extract arguments related to electric cars from the database.
This commit adds a new script, `example_fetch_docs.py`, which fetches documents from the ArangoDB database. The script uses the Ollama library to generate arguments related to electric cars from the fetched documents.
The script first queries the database to retrieve documents where the translation contains the phrase "electric car". It then iterates through the fetched documents and generates arguments using the Ollama library.
The generated arguments are based on a prompt that asks for arguments related to electric cars in the text of each document. The prompt is customized to provide a transcript of a speech given in the European Parliament.
This script will be useful for extracting arguments related to electric cars from a collection of documents.
Lasse Studion
Lasse Edfast