upload source

Dockerfile

@@ -0,0 +1,12 @@
+FROM python:3.10-slim
+
+RUN apt-get update && apt-get install -y git
+
+WORKDIR /app
+COPY . /app
+
+RUN pip install --no-cache-dir -r requirements.txt
+
+EXPOSE 7860
+
+CMD ["python", "main.py"]

main.py

@@ -0,0 +1,181 @@
+import os
+from flask import Flask, render_template, request, jsonify
+from langdetect import detect
+import torch
+import torch.nn.functional as F
+from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipeline
+
+app = Flask(__name__)
+
+# --------- Models ----------
+VI_MODEL_NAME = "wonrax/phobert-base-vietnamese-sentiment"
+EN_MODEL_NAME = "distilbert-base-uncased-finetuned-sst-2-english"
+
+device = "cuda" if torch.cuda.is_available() else "cpu"
+
+# Vietnamese model
+# vi_tokenizer = AutoTokenizer.from_pretrained(VI_MODEL_NAME, use_fast=False)
+# vi_model = AutoModelForSequenceClassification.from_pretrained(VI_MODEL_NAME).to(device)
+# vi_model.eval()
+vi_tokenizer = AutoTokenizer.from_pretrained(VI_MODEL_NAME, use_fast=False)
+vi_model = AutoModelForSequenceClassification.from_pretrained(VI_MODEL_NAME)
+vi_model.eval()
+sentiment_pipeline = pipeline("sentiment-analysis", model=vi_model, tokenizer=vi_tokenizer)
+
+
+# English model
+en_tokenizer = AutoTokenizer.from_pretrained(EN_MODEL_NAME)
+en_model = AutoModelForSequenceClassification.from_pretrained(EN_MODEL_NAME).to(device)
+en_model.eval()
+
+# Label mapping cho PhoBERT
+vi_label_map = {
+    0: ("NEGATIVE", "Tiêu cực"),
+    1: ("NEUTRAL", "Trung tính"),
+    2: ("POSITIVE", "Tích cực")
+}
+
+# Label mapping cho tiếng Anh
+en_label_map = {
+    0: ("NEGATIVE", "Negative"),
+    1: ("POSITIVE", "Positive")
+}
+
+
+# -----------------------------
+# Ngôn ngữ nhận diện
+# -----------------------------
+def detect_lang(text: str) -> str:
+    try:
+        lang = detect(text)
+        if lang.startswith("vi"):
+            return "vi"
+        elif lang.startswith("en"):
+            return "en"
+        else:
+            if any(ch in text for ch in "ăâđêôơưáàạảãấầậẩẫắằặẳẵéèẹẻẽếềệểễóòọỏõốồộổỗớờợởỡíìịỉĩúùụủũứừựửữýỳỵỷỹ"):
+                return "vi"
+            return "en"
+    except Exception:
+        if any(ch in text for ch in "ăâđêôơưáàạảãấầậẩẫắằặẳẵéèẹẻẽếềệểễóòọỏõốồộổỗớờợởỡíìịỉĩúùụủũứừựửữýỳỵỷỹ"):
+            return "vi"
+        return "en"
+
+
+# -----------------------------
+# Phân tích tiếng Việt (PhoBERT)
+# -----------------------------
+# def analyze_vi(text: str):
+#     inputs = vi_tokenizer(text, return_tensors="pt", truncation=True, padding=True).to(device)
+#     with torch.no_grad():
+#         outputs = vi_model(**inputs)
+#         logits = outputs.logits.squeeze(0)
+#         probs = torch.softmax(logits, dim=-1)
+
+#     label_idx = int(torch.argmax(probs).item())
+#     eng_label, vi_label = vi_label_map[label_idx]
+#     confidence = float(probs[label_idx].item())
+
+#     scores = {
+#         vi_label_map[i][1]: round(float(probs[i].item()), 3) for i in range(3)
+#     }
+
+#     return {
+#         "language": "vi",
+#         "label": vi_label, 
+#         "english_label": eng_label,
+#         "score": round(confidence, 3),
+#         "scores": scores
+#     }
+
+def analyze_vi(text: str):
+    if not text.strip():
+        return {"error": "Text is empty."}
+
+    # Dùng pipeline của transformers
+    result = sentiment_pipeline(text)[0]
+    label = result["label"]
+    score = round(result["score"], 3)
+
+    # Map nhãn tiếng Việt
+    label_map = {
+        "POS": "Tích cực",
+        "NEG": "Tiêu cực",
+        "NEU": "Trung tính"
+    }
+
+    vi_label = label_map.get(label, label)
+
+    # Trả kết quả tương thích với frontend
+    return {
+        "language": "vi",
+        "label": vi_label,
+        "english_label": label,  # Giữ nhãn gốc POS/NEG/NEU
+        "score": score,
+        "scores": {
+            "Tích cực": score if label == "POS" else 0.0,
+            "Trung tính": score if label == "NEU" else 0.0,
+            "Tiêu cực": score if label == "NEG" else 0.0
+        }
+    }
+# -----------------------------
+# Phân tích tiếng Anh
+# -----------------------------
+def analyze_en(text: str):
+    inputs = en_tokenizer(text, return_tensors="pt", truncation=True, padding=True).to(device)
+    with torch.no_grad():
+        outputs = en_model(**inputs)
+        logits = outputs.logits.squeeze(0)
+        probs = torch.softmax(logits, dim=-1)
+
+    label_idx = int(torch.argmax(probs).item())
+    eng_label, vi_label = en_label_map[label_idx]
+    confidence = float(probs[label_idx].item())
+
+    scores = {
+        en_label_map[i][1]: round(float(probs[i].item()), 3) for i in range(2)
+    }
+
+    return {
+        "language": "en",
+        "label": vi_label,  # Giữ English, có thể đổi sang tiếng Việt nếu muốn
+        "english_label": eng_label,
+        "score": round(confidence, 3),
+        "scores": scores
+    }
+
+
+# -----------------------------
+# Flask routes
+# -----------------------------
+@app.route("/", methods=["GET"])
+def home():
+    return render_template("index.html")
+
+
+@app.route("/analyze", methods=["POST"])
+def analyze():
+    data = request.get_json(force=True)
+    text = (data.get("text") or "").strip()
+    lang = (data.get("lang") or "auto").lower()
+    if not text:
+        return jsonify({"error": "Text is empty."}), 400
+
+    if lang == "auto":
+        lang = detect_lang(text)
+
+    if lang == "vi":
+        result = analyze_vi(text)
+    else:
+        result = analyze_en(text)
+
+    return jsonify({
+        "ok": True,
+        "input": {"text": text, "lang": lang},
+        "result": result
+    })
+
+
+if __name__ == "__main__":
+    port = int(os.environ.get("PORT", 7860))
+    app.run(host="0.0.0.0", port=port)

requirements.txt

@@ -0,0 +1,6 @@
+flask==3.0.3
+transformers==4.44.2
+torch==2.2.0+cpu
+accelerate==0.33.0
+langdetect==1.0.9
+--extra-index-url https://download.pytorch.org/whl/cpu

static/main.js

@@ -0,0 +1,88 @@
+const $ = (sel) => document.querySelector(sel);
+const analyzeBtn = $("#analyzeBtn");
+const loader = $("#loader");
+const resultBox = $("#result");
+const resLang = $("#resLang");
+const resLabel = $("#resLabel");
+const resScore = $("#resScore");
+const langSelect = $("#lang");
+const textArea = $("#text");
+
+function show(el){ el.classList.remove("hidden"); }
+function hide(el){ el.classList.add("hidden"); }
+function percent(x){ return (x*100).toFixed(1) + "%"; }
+
+function renderBars(lang, scores){
+  const keys = lang === "vi" ? ["NEGATIVE","NEUTRAL","POSITIVE"] : ["NEGATIVE","POSITIVE"];
+  keys.forEach(k=>{
+    const wrap = document.createElement("div");
+    wrap.className = "bar";
+    const fill = document.createElement("div");
+    fill.className = "bar-fill";
+    fill.style.width = "0%";
+    const label = document.createElement("div");
+    label.className = "bar-label";
+    const sc = scores[k] ?? 0;
+    label.innerHTML = `<span>${k}</span><span>${percent(sc)}</span>`;
+    wrap.appendChild(fill);
+    setTimeout(()=>{ fill.style.width = percent(sc); }, 30);
+  });
+}
+
+langSelect.addEventListener("change", (e) => {
+  const lang = e.target.value;
+  console.clear();
+  console.log("Ngôn ngữ được chọn:", lang);
+});
+
+textArea.addEventListener("input", () => {
+  console.log("Nội dung hiện tại:", textArea.value.trim());
+});
+
+analyzeBtn.addEventListener("click", async ()=>{
+  const text = textArea.value.trim();
+  const lang = langSelect.value;
+  console.log("▶️ Bắt đầu phân tích với:", { text, lang });
+
+  if (!text) {
+    alert("Nhập nội dung trước.");
+    return;
+  }
+
+  hide(resultBox);
+  show(loader);
+
+  try{
+    const r = await fetch("/analyze", {
+      method:"POST",
+      headers:{ "Content-Type":"application/json" },
+      body: JSON.stringify({ text, lang })
+    });
+    const data = await r.json();
+    hide(loader);
+
+    if (!r.ok || data.ok !== true){
+      alert(data.error || "Có lỗi xảy ra.");
+      console.error("Lỗi phản hồi:", data);
+      return;
+    }
+
+    const { input, result } = data;
+    console.log("Kết quả nhận được:", result);
+
+    resLang.textContent = input.lang.toUpperCase();
+    resLabel.textContent = result.label;
+    resLabel.style.borderColor =
+      result.label === "POSITIVE" ? "#23d2ac" :
+      result.label === "NEUTRAL"  ? "#ffd166" : "#ff6b6b";
+    resScore.textContent = percent(result.score);
+
+    renderBars(input.lang, result.scores);
+    show(resultBox);
+
+  }catch(e){
+    hide(loader);
+    alert("Lỗi mạng hoặc server.");
+    console.error("Lỗi khi gửi request:", e);
+  }
+});

static/style.css

@@ -0,0 +1,117 @@
+:root{
+  --bg:#0f1220;
+  --panel:#171a2b;
+  --panel-2:#1e2236;
+  --text:#eef1ff;
+  --muted:#9aa3c7;
+  --primary:#6c8dff;
+  --accent:#23d2ac;
+  --danger:#ff6b6b;
+  --warning:#ffd166;
+  --neutral:#a0aec0;
+  --border:#2a3150;
+  --shadow: 0 10px 25px rgba(0,0,0,.35);
+}
+
+*{box-sizing:border-box}
+html,body{height:100%}
+body{
+  margin:0;
+  display:flex;
+  font-family: system-ui, -apple-system, "Segoe UI", Roboto, Arial, "Helvetica Neue", sans-serif;
+  background: linear-gradient(180deg, #0a0d1a, #11162a);
+  color:var(--text);
+}
+
+/* Sidebar */
+.sidebar{
+  width:280px;
+  background:var(--panel);
+  border-right:1px solid var(--border);
+  display:flex; flex-direction:column;
+  position:sticky; top:0; height:100vh;
+  box-shadow: var(--shadow);
+}
+.brand{display:flex; gap:12px; justify-content: center ;padding:20px; align-items:center; border-bottom:1px solid var(--border)}
+.logo{width:40px; height:40px; border-radius:12px; display:grid; place-items:center; background:var(--panel-2); font-size:22px}
+.brand-text h1{margin:0; font-size:18px}
+.brand-text p{margin:2px 0 0; font-size:12px; color:var(--muted)}
+.nav{padding:8px}
+.nav-item{
+  display:block; padding:12px 16px; margin:6px 10px; border-radius:10px;
+  color:var(--text); text-decoration:none; background:transparent; border:1px solid transparent;
+}
+.nav-item:hover{background:var(--panel-2)}
+.nav-item.active{border-color:var(--primary); background:rgba(108,141,255,.08)}
+.nav-item.disabled{opacity:.6; cursor:not-allowed}
+.sidebar-footer{margin-top:auto; padding:16px; color:var(--muted); border-top:1px solid var(--border)}
+
+/* Main */
+.main{flex:1; display:flex; flex-direction:column; min-width:0}
+.header{
+  display:flex; align-items:center; justify-content:space-between;
+  padding:18px 24px; border-bottom:1px solid var(--border);
+  backdrop-filter: blur(6px);
+}
+.header h2{margin:0; font-size:20px}
+.badge{padding:6px 10px; background:var(--panel-2); border:1px solid var(--border); border-radius:999px; font-size:12px; color:var(--muted)}
+
+.card{
+  background:var(--panel);
+  margin:24px; padding:20px; border:1px solid var(--border); border-radius:16px;
+  box-shadow: var(--shadow);
+}
+
+label{display:block; margin:10px 0 6px; color:#e4e8ff}
+select, textarea{
+  width:100%; background:var(--panel-2); color:var(--text);
+  border:1px solid var(--border); border-radius:12px; padding:12px 14px; outline:none;
+}
+textarea{resize:vertical}
+
+.actions{display:flex; justify-content:flex-end; margin-top:12px}
+button{
+  border:1px solid transparent; background:var(--primary); color:white; padding:12px 16px;
+  border-radius:12px; cursor:pointer; font-weight:600; transition:.2s transform ease;
+  box-shadow: 0 6px 16px rgba(108,141,255,.35);
+}
+button:hover{transform: translateY(-1px)}
+button:active{transform: translateY(0)}
+
+.row{display:flex; gap:16px; margin-bottom:8px}
+.col{flex:1}
+
+/* Loader */
+.loader{display:flex; align-items:center; gap:12px; margin-top:16px; padding:12px 14px;
+  border:1px dashed var(--border); border-radius:12px; background:rgba(255,255,255,.02)}
+.spinner{
+  width:20px; height:20px; border-radius:50%;
+  border:3px solid rgba(255,255,255,.2); border-top-color:var(--accent);
+  animation:spin 0.9s linear infinite;
+}
+@keyframes spin{to{transform:rotate(360deg)}}
+.hidden{display:none}
+
+/* Result */
+.result{margin-top:16px}
+.result h3{margin:0 0 12px}
+.result-grid{display:grid; grid-template-columns: repeat(3,1fr); gap:12px}
+.result-item{background:var(--panel-2); border:1px solid var(--border); border-radius:12px; padding:12px}
+.result-item .label{display:block; font-size:12px; color:var(--muted); margin-bottom:6px}
+.value{font-weight:700}
+.badge-lg{padding:6px 10px; border-radius:10px; border:1px solid var(--border); background:rgba(255,255,255,.04)}
+.bars{margin-top:14px}
+.bar{
+  background:var(--panel-2); border:1px solid var(--border); border-radius:10px; overflow:hidden; margin:8px 0
+}
+.bar-fill{
+  height:14px; background:linear-gradient(90deg, var(--accent), var(--primary));
+  width:0%;
+  transition: width .5s ease;
+}
+.bar-label{display:flex; justify-content:space-between; font-size:12px; color:var(--muted); margin-top:4px}
+
+/* Footer */
+.footer{
+  margin: 0 24px 24px; color:var(--muted);
+}

templates/index.html

@@ -0,0 +1,87 @@
+<!DOCTYPE html>
+<html lang="vi">
+  <head>
+    <meta charset="utf-8" />
+    <meta name="viewport" content="width=device-width,initial-scale=1" />
+    <title>Sentiment Analysis (EN + VI)</title>
+    <link rel="stylesheet" href="/static/style.css" />
+  </head>
+  <body>
+    <aside class="sidebar">
+      <div class="brand">
+        <div class="brand-text">
+          <h1>Sentiment AI</h1>
+          <p>English &amp; Tiếng Việt</p>
+        </div>
+      </div>
+      <nav class="nav">
+        <a class="nav-item active" href="#">Phân tích</a>
+        <a class="nav-item disabled" href="#" title="Coming soon">Lịch sử</a>
+        <a class="nav-item disabled" href="#" title="Coming soon">Cấu hình</a>
+      </nav>
+      <footer class="sidebar-footer">
+        <small>&copy; 2025—Demo</small>
+      </footer>
+    </aside>
+
+    <main class="main">
+      <header class="header">
+        <h2>Phân tích cảm xúc</h2>
+        <div class="right">
+          <span class="badge">PhoBERT + SST-2</span>
+        </div>
+      </header>
+
+      <section class="card">
+        <div class="row">
+          <div class="col">
+            <label for="lang">Ngôn ngữ</label>
+            <select id="lang">
+              <option value="auto">Tự động</option>
+              <option value="vi" selected>Tiếng Việt</option>
+              <option value="en">English</option>
+            </select>
+          </div>
+        </div>
+
+        <label for="text">Nội dung cần phân tích</label>
+        <textarea
+          id="text"
+          rows="6"
+          placeholder="Nhập nội dung... Ví dụ: Mình rất thích sản phẩm này! / I absolutely love this!"
+        ></textarea>
+
+        <div class="actions">
+          <button id="analyzeBtn">Phân tích</button>
+        </div>
+
+        <div id="loader" class="loader hidden">
+          <div class="spinner"></div>
+          <p>Đang phân tích...</p>
+        </div>
+
+        <div id="result" class="result hidden">
+          <h3>Kết quả</h3>
+          <div class="result-grid">
+            <div class="result-item">
+              <span class="label">Ngôn ngữ</span>
+              <span id="resLang" class="value">—</span>
+            </div>
+            <div class="result-item">
+              <span class="label">Nhận định</span>
+              <span id="resLabel" class="value badge-lg">—</span>
+            </div>
+            <div class="result-item">
+              <span class="label">Độ tin cậy</span>
+              <span id="resScore" class="value">—</span>
+            </div>
+          </div>
+
+          <div id="resBars" class="bars"></div>
+        </div>
+      </section>
+    </main>
+
+    <script src="/static/main.js"></script>
+  </body>
+</html>