💡 Results: how to get indices matching original data?

[maxcorbeau]

🗺️ context: I want to use tantivy to perform a search and return a list of indices corresponding to original data I use to feed the docs, so after the search I can continue working with that original data and not with tantivy docs

So far: all the examples I have seen for retrieving the results are long the line of

query = index.parse_query("my query", ["field1", "field2"])
results = searcher.search(query, 3)
for (score, doc_address) in results.hits:
      doc = searcher.doc(best_doc_address)

with:

• doc_address of type DocAddress with .segment_ord and .doc and none of these 2 properties seem to match the original data
  • .segment_ord: I get some index which doesn't match original data nor doc_id when I provide it
  • .doc: seems always set to 0
• doc of type Document which is some custom representation of the data in tantivy format (e.g. the value for a given field is turned into a list)

🔴 Problem: none of this helps me easily work back with my original data

💡 Question: what's the easiest way from searcher.search(query, 3) to just get an ordered list of indices (from most to least relevant) corresponding to the original data I used to feed docs into tantivy?

[cjrh]

This example in the docs shows that you can obtain stored-True fields of a doc instance: https://tantivy-py.readthedocs.io/en/latest/tutorials/#building-and-executing-queries-with-the-query-parser

In the example, the title field data is obtained off the doc instance. Does this answer your question?