Hugging Face's logo

Spaces:
Lynote
/
free-ai-humanizer
Running

App Files Community
5
free-ai-humanizer
File size: 4,792 Bytes
fecd9f0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
import re

import gradio as gr


MAX_CHARS = 2_000
UTM_URL = (
    "https://lynote.ai/ai-humanizer?utm_source=huggingface"
    "&utm_medium=space&utm_campaign=hf_launch&utm_content=humanizer_lite"
)

EN_REPLACEMENTS = {
    "in today's rapidly evolving world": "today",
    "it is important to note that": "",
    "it is worth noting that": "",
    "serves as a testament to": "shows",
    "delve into": "examine",
    "leverage": "use",
    "seamless": "smooth",
    "robust": "reliable",
    "game-changer": "useful change",
    "in conclusion": "",
    "moreover": "also",
    "furthermore": "also",
}

ZH_REPLACEMENTS = {
    "值得注意的是，": "",
    "值得注意的是": "",
    "综上所述，": "",
    "综上所述": "",
    "在当今快速发展的时代": "现在",
    "赋能": "帮助",
    "助力": "帮助",
    "降本增效": "降低成本、提高效率",
    "闭环": "完整流程",
    "无缝": "顺畅",
}

PROTECTED_PATTERN = re.compile(
    r"(`[^`]+`|https?://\S+|(?:[A-Za-z]:)?(?:[/\\][\w.\-]+)+|"
    r"\b\d+(?:\.\d+)?(?:%|[A-Za-z]+)?\b|\"[^\"]*\"|'[^']*')"
)


def _protect(text: str):
    protected = []

    def replace(match):
        token = f"⟦PROTECTED_{len(protected)}⟧"
        protected.append(match.group(0))
        return token

    return PROTECTED_PATTERN.sub(replace, text), protected


def _restore(text: str, protected):
    for index, value in enumerate(protected):
        text = text.replace(f"⟦PROTECTED_{index}⟧", value)
    return text


def _replace_case_insensitive(text: str, old: str, new: str):
    return re.sub(re.escape(old), new, text, flags=re.IGNORECASE)


def humanize(text: str, voice: str):
    original = (text or "").strip()
    if not original:
        return "Please enter some text.", "No text was processed."
    if len(original) > MAX_CHARS:
        return original, f"Input is limited to {MAX_CHARS:,} characters in this demo."

    working, protected = _protect(original)
    changes = []

    for old, new in EN_REPLACEMENTS.items():
        updated = _replace_case_insensitive(working, old, new)
        if updated != working:
            changes.append(f"Replaced formulaic phrase: {old}")
            working = updated

    for old, new in ZH_REPLACEMENTS.items():
        if old in working:
            working = working.replace(old, new)
            changes.append(f"替换模板化表达：{old}")

    working = re.sub(r"[ \t]{2,}", " ", working)
    working = re.sub(r"\s+([,.;:!?，。；：！？])", r"\1", working)
    working = re.sub(r"([.!?。！？])\s*\1+", r"\1", working)
    working = re.sub(r"\n{3,}", "\n\n", working).strip(" ,，")

    if voice == "Concise":
        working = re.sub(r"\b(very|really|highly|extremely)\s+", "", working, flags=re.I)
        working = working.replace("非常", "").replace("极其", "")
    elif voice == "Professional":
        working = re.sub(r"\b(a lot of)\b", "many", working, flags=re.I)
        working = working.replace("挺多", "较多").replace("搞定", "完成")

    result = _restore(working, protected)
    if not changes and result == original:
        note = "No high-confidence template phrases were changed."
    else:
        note = f"Applied {len(changes)} conservative style edit(s). Protected numbers, URLs, paths, code, and quotes."

    return result, note


DESCRIPTION = f"""
This free, local demo removes a small set of high-confidence AI-writing clichés while
protecting numbers, URLs, paths, code, and quoted text. It is based on ideas from
[`humanize-text`](https://github.com/lynote-ai/humanize-text) and
[`humanize-text-skill`](https://github.com/lynote-ai/humanize-text-skill).

It is a writing-quality aid—not a guarantee that text will be classified as human.
[Try Lynote's full multi-stage humanizer]({UTM_URL}).
"""

with gr.Blocks(title="Lynote Humanizer Lite") as demo:
    gr.Markdown("# ✍️ Lynote Humanizer Lite")
    gr.Markdown(DESCRIPTION)
    with gr.Row():
        source = gr.Textbox(label="Draft", lines=14, max_lines=20, placeholder="Paste up to 2,000 characters…")
        output = gr.Textbox(label="Edited draft", lines=14, max_lines=20)
    voice = gr.Radio(["Natural", "Concise", "Professional"], value="Natural", label="Editing style")
    run = gr.Button("Humanize conservatively", variant="primary")
    notes = gr.Markdown()
    run.click(humanize, inputs=[source, voice], outputs=[output, notes])
    gr.Examples(
        examples=[
            ["It is worth noting that this robust tool serves as a testament to our commitment to innovation.", "Natural"],
            ["值得注意的是，我们通过这套方案赋能团队，形成降本增效闭环。", "Concise"],
        ],
        inputs=[source, voice],
    )


if __name__ == "__main__":
    demo.launch()