Adds parser package

This package is responsible for parsing the user prompts into jobs that
are carried by mather to find the best matches.

elise/src/parser/__init__.py

@@ -0,0 +1 @@
+from .parser import SentenceParser

elise/src/parser/parser.py

@@ -0,0 +1,60 @@
+from typing import List, Union
+
+import transformers
+
+
+class SentenceParser:
+    mappings = {
+        "Amenity": "Semantic",
+        "Price": "Price",
+        "Hours": "Hours",
+        "Dish": "Menu",
+        "Restaurant_Name": "Name",
+        "Location": "Location",
+        "Cuisine": "Semantic",
+    }
+
+    def __init__(self, pipe):
+        self.pipe = pipe
+
+    @classmethod
+    def from_huggingface(cls, model_path: str):
+        token_classifier = transformers.pipeline(
+            "token-classification",
+            model=model_path,
+            aggregation_strategy="simple",
+        )
+        return cls(token_classifier)
+
+    def get_ner(self, sentences: Union[str, List[str]]):
+        """Identifies the user intents from prompts
+        returns: list of dictionaries{category: words}
+        """
+        if isinstance(sentences, str):
+            sentences = [sentences]
+        # format output
+        list_out = []
+        for items in self.pipe(sentences):
+            sentence_ner = {}
+            for recognized_token in items:
+                if recognized_token["entity_group"] in list(sentence_ner.keys()):
+                    sentence_ner[recognized_token["entity_group"]].append(
+                        recognized_token["word"]
+                    )
+                else:
+                    sentence_ner[recognized_token["entity_group"]] = [
+                        recognized_token["word"]
+                    ]
+            list_out.append(sentence_ner)
+        return list_out
+
+    def get_jobs(self, ners):
+        """Maps user's intents to Matcher jobs"""
+        list_out = []
+        for item in ners:
+            jobs = {}
+            for ner, value in item.items():
+                if ner in self.mappings.keys():
+                    jobs[self.mappings[ner]] = value
+            list_out.append(jobs)
+        return list_out