Here is our Python-based project for the M2 AIC - Recherche et Extraction d'Information courses.
Based on an Elasticsearch instance, the engine is made to run temporal query on a database composed of french newpapers from 2012 to 2015. Those queries give answers as a timeline, where results are spread according to the event they refer to.
For example, if we search "séisme" (hearthquake), we should find not only the most relevant results according to a similarity with the query "séisme", but we need to obtain results corresponding to the several event according to unique hearthquakes that were dealt with in the newspapers.
The different mesures of peak detection, and a longer explanation of the project is available with the reports.
We use Python 3.
Please refer to Rapport_Installation_Elasticsearch.md,
Rapport_Python_Elasticsearch and Rapport_Indexation.md for installing and
configuring Elasticsearch, the python module elasticsearch-py and for indexing
the database.
Execute the main.py script. For more information :
python scr/main.py --help
Matthieu RÉ, Jonathan CROUZET, Kevin PASINI and Amal TARGHI