tsh-code

uvicorn[standard]
gunicorn
nltk
easynmt
sacremoses

tsh-code

FROM amazon/aws-lambda-python:3.9

COPY ./requirements.txt .

RUN yum -y install gcc-c++
RUN pip install --no-cache-dir torch --extra-index-url https://download.pytorch.org/whl/cpu
RUN pip install --no-cache-dir -r requirements.txt

ENV HOME /tmp

tsh-code

Your task is to find unique individuals in the given text. As a result, return an array of objects in the given format:

[
  {
    firstName: string,
    lastName: string,
    presumedGender: "male" | "female" | "unknown"
  }
]

tsh-code

{
   "model":"gpt-4",
   "messages":[
      {
         "role":"system",
         "content":"Your task is to find unique individuals in the given text. As a result return an array of objects in given format: [{firstName: string, lastName: string, presumedGender: 'male' | 'female' | 'unknown'}] As an answer I expect only an array of objects."
      },
      {
         "role":"user",
         "content":"${textGoesHere}"

tsh-code

[
   {
      "firstName":"Lisa",
      "lastName":"Turner",
      "presumedGender":"female"
   },
   {
      "firstName":"Mark",
      "lastName":"Davis",
      "presumedGender":"male"

tsh-code

[
  {
    "text": "Lisa Turner",
    "person": {
      "firstName": "lisa",
      "lastName": "turner",
      "honorific": "",
      "presumed_gender": "female"
    }
  },

tsh-code

[
  {
    "text": "Lisa Turner, Mark Davis",
    "person": {
      "firstName": "lisa turner",
      "lastName": "mark davis",
      "honorific": "",
      "presumed_gender": "null"
    }
  }

tsh-code

from flask import Flask, request
from datasets import load_dataset, Dataset
import json
from nltk.tokenize import sent_tokenize, word_tokenize

nlp = spacy.load("en_core_web_trf")
nlp.add_pipe("span_marker",config={"model": "lxyuan/span-marker-bert-base-multilingual-cased-multinerd"})

app = Flask(__name__)

tsh-code

import spacy

nlp = spacy.load("en_core_web_trf")
nlp.add_pipe("span_marker", config={"model": "lxyuan/span-marker-bert-base-multilingual-cased-multinerd"})

def extract_people(text: str):
   entities = nlp(text)
   full_names = set()

   for entity in entities.ents:

tsh-code

from span_marker import SpanMarkerModel
modelPreTrained = SpanMarkerModel.from_pretrained("tomaarsen/span-marker-mbert-base-multinerd")
modelPreTrained.try_cuda()

def extract_people(text:str):
   entities = modelPreTrained.predict(text)
   full_names = set()
   for entity in entities.ents:
       if entity['label'] == 'PER':
           # Check if the entity has both a first name and a last name