config.yaml

# Podcast dataset path
dataset_path: /unix/cdtdisspotify/data/spotify-podcasts-2020/

#### Feature Extraction Configuration ####
# Path to list of podcast episode uri's to extract features from
features_uri_path: ./data/uri_list_2.txt

# Output location for extracted features. This should just be the base output location
# as the extraction code creates feature specific folders within this directory (e.g. opensmile)
features_output_path: ./features

# Number of worker threads to use for feature extraction
features_num_workers: 8

#### Search Configuration ####
# Elasticsearch cluster UCL
search_es_url: http://localhost:9200/segments/_search

# Num of segments to return from the Elasticsearch query
search_es_num: 100

# Directory to cache transformer model in
search_cache_dir: ./transformers

# BERT Rerank model to use
search_rerank_model: amberoad/bert-multilingual-passage-reranking-msmarco