Closed enoriega closed 3 years ago
Can you please the information in this JSON block?
@MihaiSurdeanu sorry, I didn't follow
I don't understand the fields in the JSON example you're showing. Can you explain the new fields you added?
The block as a whole is the context frame that the FRIES output generates with context information. What I added is the freqs dictionary. In it, the key represents a context type and the value is another dictionary where the key is the sentence distance and the value the # of times it appears that far away.
For example: cl:CL:0000233 appears once in the same sentence (key 0), once one sentence away and once two sentences away. There was no species context in the neighborhood of the mention, therefore the mention got assigned the default species in the paper: taxonomy:10090 and the value of the entry is "default" instead of the frequency counter.
At the end, for every context assigned to a mention, there is going to be one entry in the freqs dictionary
Oh, I get it now. And the code looks good. merging next.
Edited the
FriesOutput
class to add the frequencies metadata as part of the context frames.An example of how it is represented in JSON:
A new JSON object is added as part of the "facets" object in the context frame. When the default species was used, I put
default
instead.