Spaces:
Paused
Paused
Upload app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,795 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import streamlit as st
|
| 2 |
+
import pandas as pd
|
| 3 |
+
import numpy as np
|
| 4 |
+
import matplotlib.pyplot as plt
|
| 5 |
+
import requests
|
| 6 |
+
import json
|
| 7 |
+
import io
|
| 8 |
+
|
| 9 |
+
# Konfiguracja
|
| 10 |
+
st.set_page_config(page_title="🤖 Agent Analityczny", layout="wide")
|
| 11 |
+
|
| 12 |
+
st.title("🤖 Agent Analityczny")
|
| 13 |
+
st.write("(Logi + wykresy)")
|
| 14 |
+
|
| 15 |
+
# Session state
|
| 16 |
+
if 'df' not in st.session_state:
|
| 17 |
+
st.session_state.df = None
|
| 18 |
+
if 'file_info' not in st.session_state:
|
| 19 |
+
st.session_state.file_info = None
|
| 20 |
+
if 'analysis_history' not in st.session_state:
|
| 21 |
+
st.session_state.analysis_history = []
|
| 22 |
+
if 'questions_count' not in st.session_state:
|
| 23 |
+
st.session_state.questions_count = 0
|
| 24 |
+
if 'quick_question' not in st.session_state:
|
| 25 |
+
st.session_state.quick_question = ""
|
| 26 |
+
|
| 27 |
+
# API
|
| 28 |
+
DEEPSEEK_API_URL = "https://api.deepseek.com/v1/chat/completions"
|
| 29 |
+
|
| 30 |
+
def smart_csv_detection(file_content):
|
| 31 |
+
"""Inteligentne wykrywanie formatu CSV"""
|
| 32 |
+
separators = [';', ',', '\t', '|']
|
| 33 |
+
encodings = ['utf-8', 'cp1250', 'latin1', 'iso-8859-1']
|
| 34 |
+
|
| 35 |
+
st.info("🔍 Analizuję format pliku...")
|
| 36 |
+
|
| 37 |
+
try:
|
| 38 |
+
sample_text = None
|
| 39 |
+
for encoding in encodings:
|
| 40 |
+
try:
|
| 41 |
+
file_content.seek(0)
|
| 42 |
+
sample_text = file_content.read(2000).decode(encoding, errors='ignore')
|
| 43 |
+
break
|
| 44 |
+
except:
|
| 45 |
+
continue
|
| 46 |
+
|
| 47 |
+
if not sample_text:
|
| 48 |
+
return None, "Nie można określić kodowania pliku"
|
| 49 |
+
|
| 50 |
+
file_content.seek(0)
|
| 51 |
+
|
| 52 |
+
except Exception as e:
|
| 53 |
+
return None, f"Błąd odczytu pliku: {e}"
|
| 54 |
+
|
| 55 |
+
results = []
|
| 56 |
+
|
| 57 |
+
for separator in separators:
|
| 58 |
+
for encoding in encodings:
|
| 59 |
+
try:
|
| 60 |
+
file_content.seek(0)
|
| 61 |
+
|
| 62 |
+
test_df = pd.read_csv(
|
| 63 |
+
file_content,
|
| 64 |
+
sep=separator,
|
| 65 |
+
encoding=encoding,
|
| 66 |
+
nrows=5,
|
| 67 |
+
quotechar='"',
|
| 68 |
+
skipinitialspace=True,
|
| 69 |
+
on_bad_lines='skip'
|
| 70 |
+
)
|
| 71 |
+
|
| 72 |
+
num_cols = len(test_df.columns)
|
| 73 |
+
quality_score = 0
|
| 74 |
+
|
| 75 |
+
if num_cols > 1:
|
| 76 |
+
quality_score += num_cols * 10
|
| 77 |
+
|
| 78 |
+
problematic_names = sum(1 for col in test_df.columns if ';' in str(col) or ',' in str(col))
|
| 79 |
+
quality_score -= problematic_names * 50
|
| 80 |
+
|
| 81 |
+
if num_cols > 1 and len(test_df) > 0:
|
| 82 |
+
first_row = test_df.iloc[0]
|
| 83 |
+
valid_cells = sum(1 for val in first_row if pd.notna(val) and str(val).strip())
|
| 84 |
+
quality_score += valid_cells * 5
|
| 85 |
+
|
| 86 |
+
results.append({
|
| 87 |
+
'separator': separator,
|
| 88 |
+
'encoding': encoding,
|
| 89 |
+
'columns': num_cols,
|
| 90 |
+
'quality': quality_score,
|
| 91 |
+
'sample_df': test_df,
|
| 92 |
+
'column_names': list(test_df.columns)
|
| 93 |
+
})
|
| 94 |
+
|
| 95 |
+
except Exception as e:
|
| 96 |
+
continue
|
| 97 |
+
|
| 98 |
+
if not results:
|
| 99 |
+
return None, "Nie udało się wykryć formatu pliku"
|
| 100 |
+
|
| 101 |
+
best_result = max(results, key=lambda x: x['quality'])
|
| 102 |
+
|
| 103 |
+
st.success(f"✅ Wykryty format:")
|
| 104 |
+
st.write(f"**Separator:** '{best_result['separator']}'")
|
| 105 |
+
st.write(f"**Kodowanie:** {best_result['encoding']}")
|
| 106 |
+
st.write(f"**Kolumny:** {best_result['columns']}")
|
| 107 |
+
st.write(f"**Jakość:** {best_result['quality']} punktów")
|
| 108 |
+
|
| 109 |
+
if len(results) > 1:
|
| 110 |
+
with st.expander("🔧 Inne możliwe formaty", expanded=False):
|
| 111 |
+
for i, result in enumerate(sorted(results, key=lambda x: x['quality'], reverse=True)[:3]):
|
| 112 |
+
if result != best_result:
|
| 113 |
+
st.write(f"Opcja {i+1}: separator='{result['separator']}', kodowanie={result['encoding']}, kolumny={result['columns']}")
|
| 114 |
+
|
| 115 |
+
return best_result, None
|
| 116 |
+
|
| 117 |
+
def load_csv_with_config(file_content, config):
|
| 118 |
+
"""Wczytaj pełny CSV z wykrytą konfiguracją"""
|
| 119 |
+
try:
|
| 120 |
+
file_content.seek(0)
|
| 121 |
+
|
| 122 |
+
df = pd.read_csv(
|
| 123 |
+
file_content,
|
| 124 |
+
sep=config['separator'],
|
| 125 |
+
encoding=config['encoding'],
|
| 126 |
+
quotechar='"',
|
| 127 |
+
skipinitialspace=True,
|
| 128 |
+
on_bad_lines='skip'
|
| 129 |
+
)
|
| 130 |
+
|
| 131 |
+
return df, None
|
| 132 |
+
|
| 133 |
+
except Exception as e:
|
| 134 |
+
return None, f"Błąd wczytywania: {e}"
|
| 135 |
+
|
| 136 |
+
def safe_display_data(df, max_rows=5):
|
| 137 |
+
"""Bezpieczne wyświetlanie danych"""
|
| 138 |
+
try:
|
| 139 |
+
if len(df) <= max_rows:
|
| 140 |
+
st.dataframe(df, use_container_width=True)
|
| 141 |
+
else:
|
| 142 |
+
st.dataframe(df.head(max_rows), use_container_width=True)
|
| 143 |
+
st.info(f"Pokazano {max_rows} z {len(df)} wierszy")
|
| 144 |
+
except:
|
| 145 |
+
st.warning("⚠️ Problem z wyświetlaniem tabeli. Pokazuję jako tekst:")
|
| 146 |
+
for i in range(min(max_rows, len(df))):
|
| 147 |
+
row_data = []
|
| 148 |
+
for col in df.columns[:6]:
|
| 149 |
+
try:
|
| 150 |
+
val = df.iloc[i][col]
|
| 151 |
+
row_data.append(f"{col}: {val}")
|
| 152 |
+
except:
|
| 153 |
+
row_data.append(f"{col}: [błąd]")
|
| 154 |
+
st.text(f"Wiersz {i+1}: {' | '.join(row_data)}")
|
| 155 |
+
|
| 156 |
+
def generate_interpretation(question, result, code, api_key):
|
| 157 |
+
"""Generuje interpretację wyników przez AI"""
|
| 158 |
+
if not api_key:
|
| 159 |
+
return "⚠️ Brak klucza API - interpretacja niedostępna"
|
| 160 |
+
|
| 161 |
+
prompt = f"""
|
| 162 |
+
Jako ekspert analizy danych, przygotuj zwięzłą interpretację wyników w języku polskim:
|
| 163 |
+
|
| 164 |
+
Pytanie użytkownika: {question}
|
| 165 |
+
Wynik analizy: {str(result)[:800] if result is not None else "Brak wyniku"}
|
| 166 |
+
Użyty kod: {code}
|
| 167 |
+
|
| 168 |
+
Przygotuj interpretację w formacie:
|
| 169 |
+
|
| 170 |
+
## 📊 Podsumowanie
|
| 171 |
+
[1-2 zdania o głównym wyniku]
|
| 172 |
+
|
| 173 |
+
## 🔍 Kluczowe wnioski
|
| 174 |
+
- [Wniosek 1]
|
| 175 |
+
- [Wniosek 2]
|
| 176 |
+
- [Wniosek 3 jeśli jest]
|
| 177 |
+
|
| 178 |
+
## 💡 Sugestie dalszych analiz
|
| 179 |
+
[Konkretne pomysły na kolejne pytania]
|
| 180 |
+
|
| 181 |
+
Bądź konkretny i praktyczny. Używaj prostego języka.
|
| 182 |
+
"""
|
| 183 |
+
|
| 184 |
+
try:
|
| 185 |
+
headers = {"Authorization": f"Bearer {api_key}", "Content-Type": "application/json"}
|
| 186 |
+
payload = {
|
| 187 |
+
"model": "deepseek-chat",
|
| 188 |
+
"messages": [{"role": "user", "content": prompt}],
|
| 189 |
+
"temperature": 0.4,
|
| 190 |
+
"max_tokens": 600
|
| 191 |
+
}
|
| 192 |
+
|
| 193 |
+
response = requests.post(DEEPSEEK_API_URL, headers=headers, json=payload, timeout=30)
|
| 194 |
+
|
| 195 |
+
if response.status_code == 200:
|
| 196 |
+
return response.json()['choices'][0]['message']['content']
|
| 197 |
+
else:
|
| 198 |
+
return f"❌ Błąd API interpretacji: {response.status_code}"
|
| 199 |
+
|
| 200 |
+
except Exception as e:
|
| 201 |
+
return f"❌ Błąd interpretacji: {str(e)}"
|
| 202 |
+
|
| 203 |
+
def simple_query_ai(prompt, api_key):
|
| 204 |
+
"""Zapytanie do AI"""
|
| 205 |
+
if not api_key:
|
| 206 |
+
return None
|
| 207 |
+
|
| 208 |
+
headers = {"Authorization": f"Bearer {api_key}", "Content-Type": "application/json"}
|
| 209 |
+
payload = {
|
| 210 |
+
"model": "deepseek-chat",
|
| 211 |
+
"messages": [{"role": "user", "content": prompt}],
|
| 212 |
+
"temperature": 0.3,
|
| 213 |
+
"max_tokens": 800
|
| 214 |
+
}
|
| 215 |
+
|
| 216 |
+
try:
|
| 217 |
+
response = requests.post(DEEPSEEK_API_URL, headers=headers, json=payload, timeout=30)
|
| 218 |
+
if response.status_code == 200:
|
| 219 |
+
return response.json()['choices'][0]['message']['content']
|
| 220 |
+
else:
|
| 221 |
+
st.error(f"API Error: {response.status_code}")
|
| 222 |
+
return None
|
| 223 |
+
except Exception as e:
|
| 224 |
+
st.error(f"Connection error: {e}")
|
| 225 |
+
return None
|
| 226 |
+
|
| 227 |
+
def extract_python_code(text):
|
| 228 |
+
"""Wyciągnij kod Python"""
|
| 229 |
+
if "```python" in text:
|
| 230 |
+
start = text.find("```python") + 9
|
| 231 |
+
end = text.find("```", start)
|
| 232 |
+
return text[start:end].strip()
|
| 233 |
+
elif "```" in text:
|
| 234 |
+
start = text.find("```") + 3
|
| 235 |
+
end = text.find("```", start)
|
| 236 |
+
return text[start:end].strip()
|
| 237 |
+
else:
|
| 238 |
+
return text.strip()
|
| 239 |
+
|
| 240 |
+
def safe_execute_code(code, df):
|
| 241 |
+
"""Wykonaj kod bezpiecznie z szczegółowymi logami"""
|
| 242 |
+
|
| 243 |
+
# Logi wykonania
|
| 244 |
+
st.subheader("🔍 Logi wykonania:")
|
| 245 |
+
log_container = st.container()
|
| 246 |
+
|
| 247 |
+
with log_container:
|
| 248 |
+
st.write("**Krok 1:** Sprawdzanie bezpieczeństwa kodu...")
|
| 249 |
+
|
| 250 |
+
# Usuń problematyczne linijki
|
| 251 |
+
problematic_patterns = ['read_csv', 'pd.read', 'pandas.read', 'to_csv', 'read_excel', 'open(', 'file(']
|
| 252 |
+
lines = code.split('\n')
|
| 253 |
+
clean_lines = []
|
| 254 |
+
removed_lines = []
|
| 255 |
+
|
| 256 |
+
for line in lines:
|
| 257 |
+
if not any(pattern in line for pattern in problematic_patterns):
|
| 258 |
+
clean_lines.append(line)
|
| 259 |
+
else:
|
| 260 |
+
removed_lines.append(line.strip())
|
| 261 |
+
|
| 262 |
+
if removed_lines:
|
| 263 |
+
st.warning(f"🛡️ Usunięto {len(removed_lines)} potencjalnie niebezpiecznych linijek:")
|
| 264 |
+
for line in removed_lines:
|
| 265 |
+
st.code(line, language='python')
|
| 266 |
+
else:
|
| 267 |
+
st.success("✅ Kod jest bezpieczny")
|
| 268 |
+
|
| 269 |
+
cleaned_code = '\n'.join(clean_lines)
|
| 270 |
+
|
| 271 |
+
st.write("**Krok 2:** Przygotowanie środowiska wykonania...")
|
| 272 |
+
|
| 273 |
+
# Bezpieczne środowisko z matplotlib
|
| 274 |
+
safe_globals = {
|
| 275 |
+
'pd': pd,
|
| 276 |
+
'np': np,
|
| 277 |
+
'plt': plt,
|
| 278 |
+
'df': df,
|
| 279 |
+
'len': len,
|
| 280 |
+
'sum': sum,
|
| 281 |
+
'max': max,
|
| 282 |
+
'min': min,
|
| 283 |
+
'abs': abs,
|
| 284 |
+
'round': round,
|
| 285 |
+
'str': str,
|
| 286 |
+
'int': int,
|
| 287 |
+
'float': float,
|
| 288 |
+
'list': list,
|
| 289 |
+
'dict': dict,
|
| 290 |
+
'sorted': sorted,
|
| 291 |
+
'enumerate': enumerate,
|
| 292 |
+
'range': range
|
| 293 |
+
}
|
| 294 |
+
|
| 295 |
+
st.success(f"✅ Dostępne moduły: pd, np, plt, df")
|
| 296 |
+
|
| 297 |
+
st.write("**Krok 3:** Wykonywanie kodu...")
|
| 298 |
+
|
| 299 |
+
safe_locals = {'result': None, 'fig': None}
|
| 300 |
+
|
| 301 |
+
try:
|
| 302 |
+
# Wykonaj cały kod naraz (bezpieczniej dla wielolinijkowych struktur)
|
| 303 |
+
progress_bar = st.progress(0)
|
| 304 |
+
progress_bar.progress(0.3)
|
| 305 |
+
|
| 306 |
+
# Spróbuj wykonać cały kod
|
| 307 |
+
exec(cleaned_code, safe_globals, safe_locals)
|
| 308 |
+
|
| 309 |
+
progress_bar.progress(1.0)
|
| 310 |
+
st.success("✅ Kod wykonany pomyślnie")
|
| 311 |
+
|
| 312 |
+
# Sprawdź co zostało utworzone
|
| 313 |
+
result = safe_locals.get('result')
|
| 314 |
+
fig = safe_locals.get('fig')
|
| 315 |
+
|
| 316 |
+
st.write("**Krok 4:** Sprawdzenie wyników...")
|
| 317 |
+
|
| 318 |
+
if result is not None:
|
| 319 |
+
result_type = type(result).__name__
|
| 320 |
+
if hasattr(result, 'shape'):
|
| 321 |
+
st.info(f"📊 Wynik: {result_type} o rozmiarze {result.shape}")
|
| 322 |
+
else:
|
| 323 |
+
st.info(f"📊 Wynik: {result_type}")
|
| 324 |
+
else:
|
| 325 |
+
st.warning("⚠️ Brak wyniku w zmiennej 'result'")
|
| 326 |
+
|
| 327 |
+
if fig is not None:
|
| 328 |
+
st.success("📈 Wykres został utworzony")
|
| 329 |
+
|
| 330 |
+
# Sprawdź czy matplotlib ma aktywne figury
|
| 331 |
+
if plt.get_fignums():
|
| 332 |
+
if fig is None:
|
| 333 |
+
fig = plt.gcf() # Pobierz aktualną figurę
|
| 334 |
+
st.info("📈 Wykrywam wykres matplotlib")
|
| 335 |
+
|
| 336 |
+
return result, fig, None
|
| 337 |
+
|
| 338 |
+
except Exception as e:
|
| 339 |
+
st.error(f"❌ Błąd wykonania: {str(e)}")
|
| 340 |
+
return None, None, str(e)
|
| 341 |
+
|
| 342 |
+
# Sidebar
|
| 343 |
+
with st.sidebar:
|
| 344 |
+
st.header("⚙️ Konfiguracja")
|
| 345 |
+
api_key = st.text_input("🔑 DeepSeek API Key", type="password")
|
| 346 |
+
|
| 347 |
+
st.header("📁 Upload pliku")
|
| 348 |
+
uploaded_file = st.file_uploader("Wybierz plik CSV", type=['csv', 'txt'])
|
| 349 |
+
|
| 350 |
+
if uploaded_file:
|
| 351 |
+
st.info("📋 Aplikacja automatycznie wykryje format pliku")
|
| 352 |
+
|
| 353 |
+
# Opcja ręcznego wyboru
|
| 354 |
+
manual_override = st.checkbox("🔧 Ręczne ustawienia (zaawansowane)")
|
| 355 |
+
|
| 356 |
+
if manual_override:
|
| 357 |
+
manual_sep = st.selectbox("Separator", [';', ',', '\t', '|'])
|
| 358 |
+
manual_enc = st.selectbox("Kodowanie", ['utf-8', 'cp1250', 'latin1'])
|
| 359 |
+
st.session_state.manual_config = {
|
| 360 |
+
'separator': manual_sep,
|
| 361 |
+
'encoding': manual_enc,
|
| 362 |
+
'columns': 0,
|
| 363 |
+
'quality': 0,
|
| 364 |
+
'sample_df': None,
|
| 365 |
+
'column_names': []
|
| 366 |
+
}
|
| 367 |
+
|
| 368 |
+
# Historia i szybkie pytania
|
| 369 |
+
if st.session_state.df is not None:
|
| 370 |
+
st.header("📊 Szybkie pytania")
|
| 371 |
+
|
| 372 |
+
# Podstawowe pytania
|
| 373 |
+
if st.button("📏 Ile wierszy?"):
|
| 374 |
+
st.session_state.quick_question = "Ile jest wierszy w danych?"
|
| 375 |
+
if st.button("📋 Jakie kolumny?"):
|
| 376 |
+
st.session_state.quick_question = "Pokaż nazwy wszystkich kolumn"
|
| 377 |
+
if st.button("🔢 Statystyki"):
|
| 378 |
+
st.session_state.quick_question = "Pokaż podstawowe statystyki numeryczne"
|
| 379 |
+
|
| 380 |
+
# Szybkie wykresy
|
| 381 |
+
st.subheader("📈 Szybkie wykresy")
|
| 382 |
+
numeric_cols = st.session_state.df.select_dtypes(include=[np.number]).columns
|
| 383 |
+
categorical_cols = st.session_state.df.select_dtypes(include=['object']).columns
|
| 384 |
+
|
| 385 |
+
if len(numeric_cols) > 0:
|
| 386 |
+
if st.button("📊 Histogram pierwszej kolumny"):
|
| 387 |
+
first_numeric = numeric_cols[0]
|
| 388 |
+
st.session_state.quick_question = f"Stwórz histogram dla kolumny {first_numeric}"
|
| 389 |
+
|
| 390 |
+
if len(categorical_cols) > 0:
|
| 391 |
+
if st.button("📈 Wykres kategorii"):
|
| 392 |
+
first_categorical = categorical_cols[0]
|
| 393 |
+
st.session_state.quick_question = f"Stwórz wykres słupkowy pokazujący ile jest każdej kategorii w kolumnie {first_categorical}"
|
| 394 |
+
|
| 395 |
+
if len(numeric_cols) >= 2:
|
| 396 |
+
if st.button("🔍 Scatter plot"):
|
| 397 |
+
col1, col2 = numeric_cols[0], numeric_cols[1]
|
| 398 |
+
st.session_state.quick_question = f"Stwórz scatter plot dla kolumn {col1} i {col2}"
|
| 399 |
+
|
| 400 |
+
# Historia
|
| 401 |
+
if st.session_state.analysis_history:
|
| 402 |
+
st.header("📜 Historia")
|
| 403 |
+
st.write(f"Wykonanych analiz: **{len(st.session_state.analysis_history)}**")
|
| 404 |
+
|
| 405 |
+
# Ostatnie pytanie z info o wykresie
|
| 406 |
+
if st.session_state.analysis_history:
|
| 407 |
+
last = st.session_state.analysis_history[-1]
|
| 408 |
+
st.write("**Ostatnie pytanie:**")
|
| 409 |
+
st.write(f"_{last['question'][:50]}..._")
|
| 410 |
+
if last.get('has_plot', False):
|
| 411 |
+
st.write("📈 *Zawierało wykres*")
|
| 412 |
+
|
| 413 |
+
# Statystyki historii
|
| 414 |
+
total_plots = sum(1 for analysis in st.session_state.analysis_history if analysis.get('has_plot', False))
|
| 415 |
+
st.write(f"**Wykresy utworzone:** {total_plots}")
|
| 416 |
+
|
| 417 |
+
# Najczęstsze słowa kluczowe
|
| 418 |
+
all_questions = [analysis['question'].lower() for analysis in st.session_state.analysis_history]
|
| 419 |
+
common_words = []
|
| 420 |
+
for question in all_questions:
|
| 421 |
+
if 'balans' in question: common_words.append('balans')
|
| 422 |
+
if 'wykres' in question or 'histogram' in question: common_words.append('wykresy')
|
| 423 |
+
if 'średnia' in question: common_words.append('średnie')
|
| 424 |
+
if 'top' in question: common_words.append('rankingi')
|
| 425 |
+
|
| 426 |
+
if common_words:
|
| 427 |
+
unique_words = list(set(common_words))
|
| 428 |
+
st.write(f"**Popularne tematy:** {', '.join(unique_words)}")
|
| 429 |
+
|
| 430 |
+
# Main interface
|
| 431 |
+
col1, col2 = st.columns([3, 1])
|
| 432 |
+
|
| 433 |
+
with col1:
|
| 434 |
+
if uploaded_file and st.button("🚀 WCZYTAJ I ANALIZUJ PLIK", type="primary"):
|
| 435 |
+
with st.spinner("🔍 Analizuję format pliku..."):
|
| 436 |
+
|
| 437 |
+
# Sprawdź czy użytkownik wybrał ręczne ustawienia
|
| 438 |
+
if hasattr(st.session_state, 'manual_config') and st.session_state.get('manual_override', False):
|
| 439 |
+
config = st.session_state.manual_config
|
| 440 |
+
st.info("🔧 Używam ręcznych ustawień")
|
| 441 |
+
else:
|
| 442 |
+
# Automatyczne wykrywanie
|
| 443 |
+
config, error = smart_csv_detection(uploaded_file)
|
| 444 |
+
|
| 445 |
+
if error:
|
| 446 |
+
st.error(f"❌ {error}")
|
| 447 |
+
st.stop()
|
| 448 |
+
|
| 449 |
+
# Wczytaj pełny plik
|
| 450 |
+
with st.spinner("📊 Wczytuję dane..."):
|
| 451 |
+
df, error = load_csv_with_config(uploaded_file, config)
|
| 452 |
+
|
| 453 |
+
if error:
|
| 454 |
+
st.error(f"❌ {error}")
|
| 455 |
+
st.stop()
|
| 456 |
+
|
| 457 |
+
st.session_state.df = df
|
| 458 |
+
st.session_state.file_info = config
|
| 459 |
+
|
| 460 |
+
# Pokaż wyniki
|
| 461 |
+
st.success(f"✅ SUKCES! Wczytano {df.shape[0]:,} wierszy × {df.shape[1]} kolumn")
|
| 462 |
+
|
| 463 |
+
# Info o kolumnach
|
| 464 |
+
st.write("📋 **Kolumny:**")
|
| 465 |
+
col_info = []
|
| 466 |
+
for i, col in enumerate(df.columns):
|
| 467 |
+
dtype = str(df[col].dtype)
|
| 468 |
+
non_null = df[col].count()
|
| 469 |
+
col_info.append(f"{i+1}. **{col}** ({dtype}) - {non_null:,}/{len(df):,} wartości")
|
| 470 |
+
|
| 471 |
+
# Pokaż kolumny w kolumnach dla czytelności
|
| 472 |
+
col_chunks = [col_info[i:i+3] for i in range(0, len(col_info), 3)]
|
| 473 |
+
for chunk in col_chunks:
|
| 474 |
+
cols = st.columns(len(chunk))
|
| 475 |
+
for i, info in enumerate(chunk):
|
| 476 |
+
cols[i].write(info)
|
| 477 |
+
|
| 478 |
+
# Podgląd danych
|
| 479 |
+
st.write("📄 **Podgląd danych:**")
|
| 480 |
+
safe_display_data(df)
|
| 481 |
+
|
| 482 |
+
# Sekcja analizy
|
| 483 |
+
if st.session_state.df is not None:
|
| 484 |
+
st.header("🔍 Analiza danych")
|
| 485 |
+
|
| 486 |
+
df = st.session_state.df
|
| 487 |
+
|
| 488 |
+
# Pokaż historię analiz jeśli są
|
| 489 |
+
if st.session_state.analysis_history:
|
| 490 |
+
with st.expander(f"📜 Historia analiz ({len(st.session_state.analysis_history)})", expanded=False):
|
| 491 |
+
for i, analysis in enumerate(reversed(st.session_state.analysis_history[-5:])): # Ostatnie 5
|
| 492 |
+
idx = len(st.session_state.analysis_history) - i
|
| 493 |
+
st.write(f"**{idx}. {analysis['question']}**")
|
| 494 |
+
|
| 495 |
+
# Ikona wykresu jeśli był
|
| 496 |
+
if analysis.get('has_plot', False):
|
| 497 |
+
st.write("📈 *Zawierał wykres*")
|
| 498 |
+
|
| 499 |
+
# Krótki wynik
|
| 500 |
+
result_preview = str(analysis['result'])[:100]
|
| 501 |
+
if len(str(analysis['result'])) > 100:
|
| 502 |
+
result_preview += "..."
|
| 503 |
+
st.write(f"Wynik: {result_preview}")
|
| 504 |
+
|
| 505 |
+
# Timestamp
|
| 506 |
+
if 'timestamp' in analysis:
|
| 507 |
+
st.caption(f"⏰ {analysis['timestamp'].strftime('%H:%M:%S')}")
|
| 508 |
+
|
| 509 |
+
st.write("---")
|
| 510 |
+
|
| 511 |
+
# Formularz do pytań
|
| 512 |
+
with st.form("analysis_form", clear_on_submit=True):
|
| 513 |
+
# Sprawdź czy jest szybkie pytanie
|
| 514 |
+
default_question = ""
|
| 515 |
+
if hasattr(st.session_state, 'quick_question') and st.session_state.quick_question:
|
| 516 |
+
default_question = st.session_state.quick_question
|
| 517 |
+
st.session_state.quick_question = ""
|
| 518 |
+
|
| 519 |
+
question = st.text_area(
|
| 520 |
+
"💬 Zadaj pytanie o dane:",
|
| 521 |
+
height=100,
|
| 522 |
+
placeholder="np. Histogram balansu, Top 10 zawodów, Rozkład wieku klientów",
|
| 523 |
+
key=f"question_input_{st.session_state.questions_count}",
|
| 524 |
+
value=default_question
|
| 525 |
+
)
|
| 526 |
+
|
| 527 |
+
col_btn1, col_btn2 = st.columns(2)
|
| 528 |
+
with col_btn1:
|
| 529 |
+
submit_button = st.form_submit_button("🧠 ANALIZUJ", type="primary")
|
| 530 |
+
with col_btn2:
|
| 531 |
+
clear_history = st.form_submit_button("🗑️ Wyczyść historię")
|
| 532 |
+
|
| 533 |
+
# Wyczyść historię jeśli kliknięto
|
| 534 |
+
if clear_history:
|
| 535 |
+
st.session_state.analysis_history = []
|
| 536 |
+
st.success("✅ Historia wyczyszczona")
|
| 537 |
+
st.rerun()
|
| 538 |
+
|
| 539 |
+
# Przetwórz pytanie
|
| 540 |
+
if submit_button and api_key and question.strip():
|
| 541 |
+
st.session_state.questions_count += 1
|
| 542 |
+
|
| 543 |
+
# Przygotuj informacje o danych
|
| 544 |
+
columns_info = f"Kolumny: {', '.join(df.columns)}"
|
| 545 |
+
dtypes_info = f"Typy danych: {dict(df.dtypes)}"
|
| 546 |
+
|
| 547 |
+
sample_data = []
|
| 548 |
+
for col in df.columns[:8]: # Pierwsze 8 kolumn
|
| 549 |
+
try:
|
| 550 |
+
if df[col].dtype in ['object', 'string']:
|
| 551 |
+
unique_vals = df[col].dropna().unique()[:3]
|
| 552 |
+
sample_data.append(f"{col}: {list(unique_vals)}")
|
| 553 |
+
else:
|
| 554 |
+
stats = f"min={df[col].min()}, max={df[col].max()}, średnia={df[col].mean():.2f}"
|
| 555 |
+
sample_data.append(f"{col}: {stats}")
|
| 556 |
+
except:
|
| 557 |
+
sample_data.append(f"{col}: [błąd próbkowania]")
|
| 558 |
+
|
| 559 |
+
sample_info = "Przykładowe dane: " + " | ".join(sample_data)
|
| 560 |
+
|
| 561 |
+
# Prompt dla AI z zachętą do tworzenia wykresów
|
| 562 |
+
prompt = f"""
|
| 563 |
+
Odpowiedz na pytanie o dane używając kodu Python. WAŻNE: Jeśli pytanie może być lepiej zobrazowane wykresem, stwórz go!
|
| 564 |
+
|
| 565 |
+
PYTANIE: {question}
|
| 566 |
+
|
| 567 |
+
INFORMACJE O DANYCH:
|
| 568 |
+
{columns_info}
|
| 569 |
+
{dtypes_info}
|
| 570 |
+
Rozmiar: {df.shape[0]} wierszy × {df.shape[1]} kolumn
|
| 571 |
+
{sample_info}
|
| 572 |
+
|
| 573 |
+
ZASADY:
|
| 574 |
+
1. DataFrame nazywa się 'df' i jest już załadowany
|
| 575 |
+
2. NIE używaj pd.read_csv(), pd.read_excel() ani podobnych
|
| 576 |
+
3. Wynik zapisz w zmiennej 'result'
|
| 577 |
+
4. Używaj pandas, numpy, matplotlib (plt)
|
| 578 |
+
5. Kod ma być prosty i skuteczny
|
| 579 |
+
6. WYKRES: Jeśli pytanie dotyczy rozkładów, trendów, porównań - ZAWSZE utwórz wykres w zmiennej 'fig'
|
| 580 |
+
|
| 581 |
+
KIEDY TWORZYĆ WYKRESY:
|
| 582 |
+
- Rozkłady wartości → histogram: plt.hist()
|
| 583 |
+
- Porównania kategorii → wykres słupkowy: plt.bar()
|
| 584 |
+
- Trendy w czasie → wykres liniowy: plt.plot()
|
| 585 |
+
- Korelacje → scatter plot: plt.scatter()
|
| 586 |
+
- Top N → wykres słupkowy
|
| 587 |
+
- Statystyki grupowe → wykres słupkowy lub pudełkowy
|
| 588 |
+
|
| 589 |
+
PRZYKŁADY KODÓW Z WYKRESAMI:
|
| 590 |
+
|
| 591 |
+
# Histogram rozkładu
|
| 592 |
+
result = df['balance'].describe()
|
| 593 |
+
fig, ax = plt.subplots(figsize=(10, 6))
|
| 594 |
+
ax.hist(df['balance'], bins=30, alpha=0.7)
|
| 595 |
+
ax.set_title('Rozkład balansu klientów')
|
| 596 |
+
ax.set_xlabel('Balans')
|
| 597 |
+
ax.set_ylabel('Liczba klientów')
|
| 598 |
+
|
| 599 |
+
# Wykres słupkowy dla kategorii
|
| 600 |
+
result = df['job'].value_counts().head(10)
|
| 601 |
+
fig, ax = plt.subplots(figsize=(12, 6))
|
| 602 |
+
result.plot(kind='bar', ax=ax)
|
| 603 |
+
ax.set_title('Top 10 zawodów')
|
| 604 |
+
ax.set_xlabel('Zawód')
|
| 605 |
+
ax.set_ylabel('Liczba osób')
|
| 606 |
+
plt.xticks(rotation=45)
|
| 607 |
+
|
| 608 |
+
# Porównanie grup
|
| 609 |
+
result = df.groupby('job')['balance'].mean().head(10)
|
| 610 |
+
fig, ax = plt.subplots(figsize=(12, 6))
|
| 611 |
+
result.plot(kind='bar', ax=ax)
|
| 612 |
+
ax.set_title('Średni balans według zawodu')
|
| 613 |
+
ax.set_ylabel('Średni balans')
|
| 614 |
+
plt.xticks(rotation=45)
|
| 615 |
+
|
| 616 |
+
PRZYKŁADY BEZ WYKRESÓW (tylko liczby):
|
| 617 |
+
- Liczba wierszy: result = len(df)
|
| 618 |
+
- Średnia: result = df['kolumna'].mean()
|
| 619 |
+
- Filtrowanie: result = len(df[df['kolumna'] > 100])
|
| 620 |
+
|
| 621 |
+
Zwróć TYLKO kod Python (bez wyjaśnień):
|
| 622 |
+
"""
|
| 623 |
+
|
| 624 |
+
# Zapytaj AI
|
| 625 |
+
with st.spinner("🤖 AI generuje kod..."):
|
| 626 |
+
st.info("🔄 Wysyłam zapytanie do DeepSeek API...")
|
| 627 |
+
ai_response = simple_query_ai(prompt, api_key)
|
| 628 |
+
|
| 629 |
+
if ai_response:
|
| 630 |
+
st.success("✅ Otrzymano odpowiedź od AI")
|
| 631 |
+
code = extract_python_code(ai_response)
|
| 632 |
+
|
| 633 |
+
st.subheader("🔧 Wygenerowany kod:")
|
| 634 |
+
st.code(code, language='python')
|
| 635 |
+
|
| 636 |
+
# Wykonaj z logowaniem
|
| 637 |
+
with st.spinner("⚡ Wykonuję analizę..."):
|
| 638 |
+
result, fig, error = safe_execute_code(code, df)
|
| 639 |
+
|
| 640 |
+
# Wyniki
|
| 641 |
+
if error:
|
| 642 |
+
st.error(f"❌ Błąd końcowy: {error}")
|
| 643 |
+
st.write("**Możliwe rozwiązania:**")
|
| 644 |
+
st.write("- Sprawdź nazwy kolumn")
|
| 645 |
+
st.write("- Upewnij się że kolumna zawiera dane liczbowe")
|
| 646 |
+
st.write("- Spróbuj prostsze pytanie")
|
| 647 |
+
else:
|
| 648 |
+
st.subheader("📊 WYNIK:")
|
| 649 |
+
if result is not None:
|
| 650 |
+
if isinstance(result, (int, float)):
|
| 651 |
+
st.metric("Wynik", f"{result:,}")
|
| 652 |
+
elif isinstance(result, str):
|
| 653 |
+
st.write(f"**{result}**")
|
| 654 |
+
elif isinstance(result, pd.DataFrame):
|
| 655 |
+
st.write("**Tabela wyników:**")
|
| 656 |
+
safe_display_data(result, max_rows=20)
|
| 657 |
+
elif isinstance(result, pd.Series):
|
| 658 |
+
st.write("**Serie danych:**")
|
| 659 |
+
safe_display_data(result.to_frame(), max_rows=20)
|
| 660 |
+
else:
|
| 661 |
+
st.write("**Wynik:**")
|
| 662 |
+
st.text(str(result)[:1000])
|
| 663 |
+
|
| 664 |
+
# WYKRESY - ulepszone wyświetlanie
|
| 665 |
+
if fig is not None:
|
| 666 |
+
st.subheader("📈 Wykres:")
|
| 667 |
+
try:
|
| 668 |
+
plt.tight_layout()
|
| 669 |
+
st.pyplot(fig, use_container_width=True)
|
| 670 |
+
st.success("✅ Wykres wyświetlony pomyślnie")
|
| 671 |
+
except Exception as plot_error:
|
| 672 |
+
st.error(f"❌ Błąd wyświetlania wykresu: {plot_error}")
|
| 673 |
+
finally:
|
| 674 |
+
plt.close(fig)
|
| 675 |
+
elif any(word in question.lower() for word in ['wykres', 'histogram', 'rozkład', 'porównaj', 'pokaż', 'wizualiz']):
|
| 676 |
+
st.info("💡 Zapytanie sugeruje wykres, ale AI go nie utworzyło. Spróbuj poprosić bezpośrednio o wykres.")
|
| 677 |
+
|
| 678 |
+
# INTERPRETACJA AI
|
| 679 |
+
st.subheader("🧠 Interpretacja AI:")
|
| 680 |
+
with st.spinner("🔮 AI interpretuje wyniki..."):
|
| 681 |
+
interpretation = generate_interpretation(question, result, code, api_key)
|
| 682 |
+
st.markdown(interpretation)
|
| 683 |
+
|
| 684 |
+
# Zapisz do historii
|
| 685 |
+
analysis_record = {
|
| 686 |
+
'question': question,
|
| 687 |
+
'code': code,
|
| 688 |
+
'result': result,
|
| 689 |
+
'interpretation': interpretation,
|
| 690 |
+
'timestamp': pd.Timestamp.now(),
|
| 691 |
+
'has_plot': fig is not None
|
| 692 |
+
}
|
| 693 |
+
st.session_state.analysis_history.append(analysis_record)
|
| 694 |
+
|
| 695 |
+
# Zachęć do kolejnych pytań
|
| 696 |
+
st.markdown("---")
|
| 697 |
+
st.success("✅ Analiza zakończona! Możesz zadać kolejne pytanie powyżej.")
|
| 698 |
+
|
| 699 |
+
# Sugestie kolejnych pytań na podstawie wyniku
|
| 700 |
+
if isinstance(result, pd.DataFrame) and len(result) > 1:
|
| 701 |
+
st.info("💡 Sugestia: Możesz zapytać o wykres dla tych wyników")
|
| 702 |
+
elif isinstance(result, (int, float)) and 'balans' in question.lower():
|
| 703 |
+
st.info("💡 Sugestia: Spróbuj 'Pokaż histogram balansu' lub 'Rozkład balansu klientów'")
|
| 704 |
+
elif 'top' in question.lower() or 'najwyższ' in question.lower():
|
| 705 |
+
st.info("💡 Sugestia: Możesz poprosić o wykres słupkowy dla tych wyników")
|
| 706 |
+
|
| 707 |
+
else:
|
| 708 |
+
st.error("❌ Nie udało się uzyskać odpowiedzi od AI")
|
| 709 |
+
st.write("**Możliwe przyczyny:**")
|
| 710 |
+
st.write("- Błąd klucza API")
|
| 711 |
+
st.write("- Problem z połączeniem internetowym")
|
| 712 |
+
st.write("- Przeciążenie serwera DeepSeek")
|
| 713 |
+
|
| 714 |
+
elif submit_button and not api_key:
|
| 715 |
+
st.warning("⚠️ Wprowadź klucz API DeepSeek w lewym panelu")
|
| 716 |
+
elif submit_button and not question.strip():
|
| 717 |
+
st.warning("⚠️ Wprowadź pytanie do analizy")
|
| 718 |
+
|
| 719 |
+
with col2:
|
| 720 |
+
st.header("📊 Informacje")
|
| 721 |
+
|
| 722 |
+
if st.session_state.df is not None:
|
| 723 |
+
df = st.session_state.df
|
| 724 |
+
|
| 725 |
+
# Podstawowe metryki
|
| 726 |
+
st.metric("📊 Wiersze", f"{len(df):,}")
|
| 727 |
+
st.metric("📋 Kolumny", len(df.columns))
|
| 728 |
+
|
| 729 |
+
# Informacje o pliku
|
| 730 |
+
if st.session_state.file_info:
|
| 731 |
+
st.subheader("📁 Format pliku")
|
| 732 |
+
info = st.session_state.file_info
|
| 733 |
+
st.write(f"**Separator:** `{info['separator']}`")
|
| 734 |
+
st.write(f"**Kodowanie:** {info['encoding']}")
|
| 735 |
+
|
| 736 |
+
# Typy danych
|
| 737 |
+
st.subheader("🏷️ Typy kolumn")
|
| 738 |
+
type_counts = df.dtypes.value_counts()
|
| 739 |
+
for dtype, count in type_counts.items():
|
| 740 |
+
st.write(f"**{str(dtype)}:** {count}")
|
| 741 |
+
|
| 742 |
+
# Braki danych
|
| 743 |
+
missing_data = df.isnull().sum()
|
| 744 |
+
missing_cols = missing_data[missing_data > 0]
|
| 745 |
+
|
| 746 |
+
if len(missing_cols) > 0:
|
| 747 |
+
st.subheader("⚠️ Braki danych")
|
| 748 |
+
for col, missing in missing_cols.items():
|
| 749 |
+
pct = (missing / len(df)) * 100
|
| 750 |
+
st.write(f"**{col}:** {missing:,} ({pct:.1f}%)")
|
| 751 |
+
else:
|
| 752 |
+
st.success("✅ Brak braków danych")
|
| 753 |
+
|
| 754 |
+
# Kolumny numeryczne - podstawowe statystyki
|
| 755 |
+
numeric_cols = df.select_dtypes(include=[np.number]).columns
|
| 756 |
+
if len(numeric_cols) > 0:
|
| 757 |
+
st.subheader("🔢 Statystyki numeryczne")
|
| 758 |
+
for col in numeric_cols[:5]:
|
| 759 |
+
try:
|
| 760 |
+
mean_val = df[col].mean()
|
| 761 |
+
min_val = df[col].min()
|
| 762 |
+
max_val = df[col].max()
|
| 763 |
+
st.write(f"**{col}:**")
|
| 764 |
+
st.write(f" Średnia: {mean_val:.2f}")
|
| 765 |
+
st.write(f" Zakres: {min_val:.2f} - {max_val:.2f}")
|
| 766 |
+
except:
|
| 767 |
+
st.write(f"**{col}:** błąd obliczeń")
|
| 768 |
+
else:
|
| 769 |
+
st.info("👆 Wczytaj plik CSV aby zobaczyć statystyki")
|
| 770 |
+
|
| 771 |
+
# Footer
|
| 772 |
+
st.markdown("---")
|
| 773 |
+
st.markdown("🔧 **Funkcje:**")
|
| 774 |
+
col_f1, col_f2 = st.columns(2)
|
| 775 |
+
|
| 776 |
+
with col_f1:
|
| 777 |
+
st.markdown("• **Automatyczne wykrywanie** formatu CSV")
|
| 778 |
+
st.markdown("• **Obsługa separatorów:** `;` `,` `tab` `|`")
|
| 779 |
+
st.markdown("• **Różne kodowania:** UTF-8, CP1250, Latin1")
|
| 780 |
+
st.markdown("• **Inteligentne wyświetlanie** (omija błędy)")
|
| 781 |
+
st.markdown("• **Szczegółowe logi** wykonania kodu")
|
| 782 |
+
|
| 783 |
+
with col_f2:
|
| 784 |
+
st.markdown("• **Interpretacja AI** wyników analizy")
|
| 785 |
+
st.markdown("• **Historia pytań** w sesji")
|
| 786 |
+
st.markdown("• **Automatyczne wykresy** (histogram, bar, scatter)")
|
| 787 |
+
st.markdown("• **Szybkie wykresy** jednym kliknięciem")
|
| 788 |
+
st.markdown("• **Ciągłość analizy** - kolejne pytania")
|
| 789 |
+
|
| 790 |
+
st.info("💡 **Tip:** Pytaj o 'histogram', 'wykres', 'rozkład', 'porównanie' - AI automatycznie utworzy odpowiedni wykres!")
|
| 791 |
+
|
| 792 |
+
# Debug info
|
| 793 |
+
if st.session_state.analysis_history:
|
| 794 |
+
plots_count = sum(1 for analysis in st.session_state.analysis_history if analysis.get('has_plot', False))
|
| 795 |
+
st.caption(f"🎯 W tej sesji: {len(st.session_state.analysis_history)} analiz, {plots_count} wykresów")
|