app.py

import gradio as gr

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

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
    
def word2vec(word):
    if not word in vectors:
        return None
    return list(vectors[word])
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
def paragraph2word(paragraph):
    tokens = tokenize(paragraph)
    mergedVector = []

    # Merge vectors
    for token in tokens:
        vector = word2vec(token)
        if len(mergedVector) == 0:
            mergedVector = vector
        else:
            for i in range(len(vector)):
                mergedVector[i] += vector[i]
    
    # Normalize
    for i in range(len(mergedVector)):
        mergedVector[i] /= len(tokens)
    
    return mergedVector

demo = gr.Interface(fn=paragraph2word, inputs="text", outputs="text")
demo.launch()