Adicionados app.py e requirements.txt; modificado README.md

README.md

@@ -1,13 +1,50 @@
 ---
-title: Sentiment Analysis Committee
-emoji: 📉
-colorFrom: green
-colorTo: red
+title: sentiment-analysis-committee
+emoji: 👥
+colorFrom: blue
+colorTo: green
 sdk: gradio
-sdk_version: 4.12.0
+sdk_version: "4.12.0"
 app_file: app.py
 pinned: false
-license: ecl-2.0
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+
+# Sentiment Analysis Committee
+
+A comprehensive sentiment analysis tool using multiple methods, including BERT (Base and Large), DistilBERT, SiEBERT, TextBlob, VADER, and AFINN.
+
+## How to Use
+
+Enter text into the interface to receive sentiment analyses from various methods. The committee's decision is based on the majority of votes among the methods.
+
+## Technical Details
+
+This project leverages various natural language processing models to evaluate the sentiment of entered text:
+
+- **BERT Base and BERT Large**: Transformer-based models providing sentiment scores and labels. BERT Large is a larger variant of BERT with more layers, potentially offering more nuanced sentiment analysis.
+- **DistilBERT**: A distilled version of BERT, optimized for speed and efficiency.
+- **SiEBERT**: A RoBERTa-based model fine-tuned for sentiment analysis.
+- **TextBlob**: Utilizes Naive Bayes classifiers, offering straightforward sentiment evaluations.
+- **VADER**: Designed for social media and short texts, giving a compound sentiment score.
+- **AFINN**: A lexical method assigning scores to words, indicating sentiment intensity.
+
+The final decision of the committee is determined by a majority vote approach, providing a balanced sentiment analysis.
+
+## Additional Information
+
+- Developed by Ramon Mayor Martins (2023)
+- E-mail: [rmayormartins@gmail.com](mailto:rmayormartins@gmail.com)
+- Homepage: [https://rmayormartins.github.io/](https://rmayormartins.github.io/)
+- Twitter: [@rmayormartins](https://twitter.com/rmayormartins)
+- GitHub: [https://github.com/rmayormartins](https://github.com/rmayormartins)
+
+## Notes
+
+- The committee's decision is democratic, based on the majority vote from the utilized methods.
+- The project is implemented in Python and hosted on Hugging Face Spaces.
+
+
+
+
+

app.py

@@ -0,0 +1,107 @@
+from transformers import pipeline
+import gradio as gr
+from textblob import TextBlob
+import numpy as np
+import nltk
+from nltk.sentiment import SentimentIntensityAnalyzer
+from afinn import Afinn
+
+
+#VADER e AFINN
+nltk.download('vader_lexicon')
+vader = SentimentIntensityAnalyzer()
+afinn = Afinn()
+
+#Hugging Face
+bert_model = pipeline("sentiment-analysis", model="bert-base-uncased")
+#BERT Large
+bert_large_model = pipeline("sentiment-analysis", model="bert-large-uncased")
+distilbert_model = pipeline("sentiment-analysis", model="distilbert-base-uncased")
+siebert_model = pipeline("sentiment-analysis", model="siebert/sentiment-roberta-large-english")
+
+
+def normalize_score(score, range_min, range_max):
+    return (score - range_min) / (range_max - range_min)
+
+
+def analyze_with_bert(text):
+    analysis = bert_model(text)
+    label, score = map_label(analysis[0]['label']), analysis[0]['score']
+    return label, score
+
+
+def analyze_with_bert_large(text):
+    analysis = bert_large_model(text)
+    label, score = map_label(analysis[0]['label']), analysis[0]['score']
+    return label, score
+
+def analyze_with_distilbert(text):
+    analysis = distilbert_model(text)
+    label, score = map_label(analysis[0]['label']), analysis[0]['score']
+    return label, score
+
+def analyze_with_siebert(text):
+    analysis = siebert_model(text)
+    return analysis[0]['label'], analysis[0]['score']
+
+def analyze_with_textblob(text):
+    analysis = TextBlob(text).sentiment
+    label = "POSITIVE" if analysis.polarity > 0 else "NEGATIVE" if analysis.polarity < 0 else "NEUTRAL"
+    normalized_score = normalize_score(analysis.polarity, -1, 1)
+    return label, normalized_score
+
+def analyze_with_vader(text):
+    scores = vader.polarity_scores(text)
+    label = "POSITIVE" if scores['compound'] > 0.05 else "NEGATIVE" if scores['compound'] < -0.05 else "NEUTRAL"
+    normalized_score = normalize_score(scores['compound'], -1, 1)
+    return label, normalized_score
+
+def analyze_with_afinn(text):
+    score = afinn.score(text)
+    label = "POSITIVE" if score > 0 else "NEGATIVE" if score < 0 else "NEUTRAL"
+    normalized_score = normalize_score(score, -5, 5)
+    return label, normalized_score
+
+#mapeio BERT e DistilBERT
+def map_label(label):
+    if label == "LABEL_0":
+        return "NEGATIVE"
+    elif label == "LABEL_1":
+        return "POSITIVE"
+    else:
+        return "NEUTRAL"
+
+
+#Comite
+def calculate_committee_decision(results):
+    #coto voto
+    vote_count = {"POSITIVE": 0, "NEGATIVE": 0, "NEUTRAL": 0}
+    for label, score in results.values():
+        vote_count[label] += 1
+
+    #maioria dos votos
+    final_label = max(vote_count, key=vote_count.get)
+    return final_label, vote_count[final_label] / len(results)
+
+
+
+
+def analyze_text(text):
+    results = {
+        "BERT Base": analyze_with_bert(text),
+        "BERT Large": analyze_with_bert_large(text),
+        "DistilBERT": analyze_with_distilbert(text),
+        "SiEBERT": analyze_with_siebert(text),
+        "TextBlob": analyze_with_textblob(text),
+        "VADER": analyze_with_vader(text),
+        "AFINN": analyze_with_afinn(text)
+    }
+
+    final_label, vote_ratio = calculate_committee_decision(results)
+    results["Committee Decision"] = {"label": final_label, "vote_ratio": vote_ratio}
+    return results
+
+
+#Gradio
+iface = gr.Interface(fn=analyze_text, inputs="text", outputs="json")
+iface.launch(debug=True)

requirements.txt

@@ -0,0 +1,5 @@
+transformers
+gradio
+textblob
+nltk
+afinn