Since the implementation of Semantic Table Retrieval seems to get rather complex I've decide to split it up in to the appropriate steps. Therefore I've opened this PR with the initial step of STR: extraction.
Missing is the MLM approach to perform the ranking of the extracted entities and select a top 10. For this we will await the response of the professor.
It also includes an API system with (extremely simple) caching for the DBpedia api, however in some small tests I did this did not seem to result in a lot of entities being detected, so perhaps some additional step is required that I cannot find in the paper.
Since the implementation of Semantic Table Retrieval seems to get rather complex I've decide to split it up in to the appropriate steps. Therefore I've opened this PR with the initial step of STR: extraction.
Missing is the MLM approach to perform the ranking of the extracted entities and select a top 10. For this we will await the response of the professor.
It also includes an API system with (extremely simple) caching for the DBpedia api, however in some small tests I did this did not seem to result in a lot of entities being detected, so perhaps some additional step is required that I cannot find in the paper.