app.py

import gradio as gr
import io
import numpy as np

# Trie
class TrieNode:
    def __init__(self):
        self.children = {}
        self.is_end_of_token = False
class Trie:
    def __init__(self):
        self.root = TrieNode()

    def insert(self, token):
        node = self.root
        for char in token:
            if char not in node.children:
                node.children[char] = TrieNode()
            node = node.children[char]
        node.is_end_of_token = True

    def search_longest_prefix(self, text, start):
        node = self.root
        longest_match = None
        current_pos = start
        
        while current_pos < len(text) and text[current_pos] in node.children:
            node = node.children[text[current_pos]]
            if node.is_end_of_token:
                longest_match = current_pos
            current_pos += 1
        
        return longest_match

# Vector Loader   
def load_vectors(fname):
    fin = io.open(fname, 'r', encoding='utf-8', newline='\n', errors='ignore')
    data = {}
    for line in fin:
        tokens = line.rstrip().split(' ')
        data[tokens[0]] = np.array(list(map(float, tokens[1:])))  # Convert to NumPy array
    del fin
    return data, sorted(data.keys(), key=len, reverse=True)
vectors, sorted_vector = load_vectors('wiki-news-300d-1M.vec')

# Tokenizer
def tokenize(text):
    trie = Trie()
    for token in sorted_vector:
        trie.insert(token)

    result = []
    start = 0
    
    while start < len(text):
        longest_match = trie.search_longest_prefix(text, start)
        if longest_match is not None:
            result.append(text[start:longest_match+1])
            start = longest_match + 1
        else:
            start += 1
    
    return result

# Interface
def onInput(paragraph, progress = gr.Progress()):
    progress(0, "Tokenizing...")
    tokens = tokenize(paragraph)
    
    progress(0.1, "Initializing merged vector...")
    if not tokens:  # Handle case with no tokens found
        return np.zeros(300).tolist()  # Return a zero vector of appropriate dimension
    
    merged_vector = np.zeros(300)  # Assuming vectors are 300-dimensional
    
    # Merge vectors using NumPy
    totalTokens = len(tokens)
    for ind, token in enumerate(tokens):
        completion = 0.7*((ind+1)/totalTokens)
        progress(0.1 + completion, f"Merging {token}, Token #{tokens.index(token)+1}/{len(tokens)}")

        vector = vectors[token]
        merged_vector += vector
    
    # Normalize
    progress(0.9, "Normalizing...")
    merged_vector /= len(tokens)
    
    progress(1, "Converting to list...")
    return merged_vector.tolist()  # Convert back to list for output

demo = gr.Interface(fn=onInput, inputs="text", outputs="text")
demo.launch()