Parser for Constituency Tree (Linguistics Analysis API)

问题

I'm using Microsofts Linguistics API. I'm trying to extract specific tokens from the returned tree. I dont see any type of parser for traversing through the tree in any documentation...

One approach I considered was to use the Stanford NLP parser but It seems a little overkill for what I need.

Is there an existing parser that I could use?

here is sample data that is returned. for example what can I use to extract "NNP" (Tom)

[{
"analyzerId": "4FA79AF1-F22C-408D-98BB-B7D7AEEF7F04", 
"result": [ ["NNP",",","NNP","."], ["WRB","VBP","PRP","NN","."] ]  },

{
"analyzerId": "22A6B758-420F-4745-8A3C-46835A67C0D2", 
"result":["(TOP (S (NNP Hi) (, ,) (NNP Tom) (. !)))","(TOP (SBARQ (WHADVP (WRB How)) (SQ (VP (VBP are)) (NP (PRP you)) (NN today) (. ?))))"]  }]

回答1:

Find my parser-to-tree (and roundtripping) source code at:

https://github.com/BSalita/Woundify/blob/master/WoundifyShared/ParseHelpers.cs

ParseHelpers is a file from the Woundify project. One of the functions of the project is to demonstrate the calling and consumption of APIs from all leading AI services providers (Microsoft, Google, HPE, IBM, Wit, Hound, etc.).

I've edited in a usage fragment from command.cs file:

foreach (Newtonsoft.Json.Linq.JToken s in arrayOfResults)
{
    ConstituencyTreeNode root = ParseHelpers.ConstituencyTreeFromText(s.ToString());
    text = ParseHelpers.TextFromConstituencyTree(root); // roundtrip
    if (text != s.ToString()) // original and roundtrip must compare equal
        throw new FormatException();
    string words = ParseHelpers.WordsFromConstituencyTree(root);
    string[] printLines = ParseHelpers.FormatConstituencyTree(root);
    foreach (string p in printLines)
        Console.WriteLine(p);
}

来源：https://stackoverflow.com/questions/41518425/parser-for-constituency-tree-linguistics-analysis-api

标签

microsoft-cognitive