Update app.py
Browse files
app.py
CHANGED
@@ -26,12 +26,27 @@ import uuid # For generating unique report IDs
|
|
26 |
# ------------------------------
|
27 |
class GroqLLM:
|
28 |
"""Enhanced LLM interface with support for generating natural language summaries."""
|
29 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
30 |
self.client = Groq(api_key=os.environ.get("GROQ_API_KEY"))
|
31 |
self.model_name = model_name
|
32 |
|
33 |
def __call__(self, prompt: Union[str, dict, List[Dict]]) -> str:
|
34 |
-
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
35 |
try:
|
36 |
# Handle different prompt formats
|
37 |
if isinstance(prompt, (dict, list)):
|
@@ -63,18 +78,39 @@ class GroqLLM:
|
|
63 |
# ------------------------------
|
64 |
class DataAnalysisAgent(CodeAgent):
|
65 |
"""Extended CodeAgent with dataset awareness and predictive analytics capabilities."""
|
|
|
66 |
def __init__(self, dataset: pd.DataFrame, *args, **kwargs):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
67 |
super().__init__(*args, **kwargs)
|
68 |
self._dataset = dataset
|
69 |
self.models = {} # To store trained models
|
70 |
|
71 |
@property
|
72 |
def dataset(self) -> pd.DataFrame:
|
73 |
-
"""Access the stored dataset
|
|
|
|
|
|
|
|
|
74 |
return self._dataset
|
75 |
|
76 |
def run(self, prompt: str) -> str:
|
77 |
-
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
78 |
dataset_info = f"""
|
79 |
Dataset Shape: {self.dataset.shape}
|
80 |
Columns: {', '.join(self.dataset.columns)}
|
@@ -96,7 +132,22 @@ class DataAnalysisAgent(CodeAgent):
|
|
96 |
|
97 |
@tool
|
98 |
def analyze_basic_stats(data: pd.DataFrame) -> str:
|
99 |
-
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
100 |
if data is None:
|
101 |
data = tool.agent.dataset
|
102 |
|
@@ -134,7 +185,21 @@ def analyze_basic_stats(data: pd.DataFrame) -> str:
|
|
134 |
|
135 |
@tool
|
136 |
def generate_correlation_matrix(data: pd.DataFrame) -> str:
|
137 |
-
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
138 |
if data is None:
|
139 |
data = tool.agent.dataset
|
140 |
|
@@ -156,7 +221,21 @@ def generate_correlation_matrix(data: pd.DataFrame) -> str:
|
|
156 |
|
157 |
@tool
|
158 |
def analyze_categorical_columns(data: pd.DataFrame) -> str:
|
159 |
-
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
160 |
if data is None:
|
161 |
data = tool.agent.dataset
|
162 |
|
@@ -197,7 +276,20 @@ def analyze_categorical_columns(data: pd.DataFrame) -> str:
|
|
197 |
|
198 |
@tool
|
199 |
def suggest_features(data: pd.DataFrame) -> str:
|
200 |
-
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
201 |
if data is None:
|
202 |
data = tool.agent.dataset
|
203 |
|
@@ -231,7 +323,21 @@ def suggest_features(data: pd.DataFrame) -> str:
|
|
231 |
|
232 |
@tool
|
233 |
def predictive_analysis(data: pd.DataFrame, target: str) -> str:
|
234 |
-
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
235 |
if data is None:
|
236 |
data = tool.agent.dataset
|
237 |
|
@@ -326,7 +432,19 @@ def predictive_analysis(data: pd.DataFrame, target: str) -> str:
|
|
326 |
# Report Exporting Function
|
327 |
# ------------------------------
|
328 |
def export_report(content: str, filename: str):
|
329 |
-
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
330 |
# Save content to a temporary HTML file
|
331 |
with tempfile.NamedTemporaryFile(delete=False, suffix='.html') as tmp_file:
|
332 |
tmp_file.write(content.encode('utf-8'))
|
|
|
26 |
# ------------------------------
|
27 |
class GroqLLM:
|
28 |
"""Enhanced LLM interface with support for generating natural language summaries."""
|
29 |
+
|
30 |
+
def __init__(self, model_name: str = "llama-3.1-8B-Instant"):
|
31 |
+
"""
|
32 |
+
Initialize the GroqLLM with a specified model.
|
33 |
+
|
34 |
+
Args:
|
35 |
+
model_name (str): The name of the language model to use.
|
36 |
+
"""
|
37 |
self.client = Groq(api_key=os.environ.get("GROQ_API_KEY"))
|
38 |
self.model_name = model_name
|
39 |
|
40 |
def __call__(self, prompt: Union[str, dict, List[Dict]]) -> str:
|
41 |
+
"""
|
42 |
+
Make the class callable as required by smolagents.
|
43 |
+
|
44 |
+
Args:
|
45 |
+
prompt (Union[str, dict, List[Dict]]): The input prompt for the language model.
|
46 |
+
|
47 |
+
Returns:
|
48 |
+
str: The generated response from the language model.
|
49 |
+
"""
|
50 |
try:
|
51 |
# Handle different prompt formats
|
52 |
if isinstance(prompt, (dict, list)):
|
|
|
78 |
# ------------------------------
|
79 |
class DataAnalysisAgent(CodeAgent):
|
80 |
"""Extended CodeAgent with dataset awareness and predictive analytics capabilities."""
|
81 |
+
|
82 |
def __init__(self, dataset: pd.DataFrame, *args, **kwargs):
|
83 |
+
"""
|
84 |
+
Initialize the DataAnalysisAgent with the provided dataset.
|
85 |
+
|
86 |
+
Args:
|
87 |
+
dataset (pd.DataFrame): The dataset to analyze.
|
88 |
+
*args: Variable length argument list.
|
89 |
+
**kwargs: Arbitrary keyword arguments.
|
90 |
+
"""
|
91 |
super().__init__(*args, **kwargs)
|
92 |
self._dataset = dataset
|
93 |
self.models = {} # To store trained models
|
94 |
|
95 |
@property
|
96 |
def dataset(self) -> pd.DataFrame:
|
97 |
+
"""Access the stored dataset.
|
98 |
+
|
99 |
+
Returns:
|
100 |
+
pd.DataFrame: The dataset stored in the agent.
|
101 |
+
"""
|
102 |
return self._dataset
|
103 |
|
104 |
def run(self, prompt: str) -> str:
|
105 |
+
"""
|
106 |
+
Override the run method to include dataset context and support predictive tasks.
|
107 |
+
|
108 |
+
Args:
|
109 |
+
prompt (str): The task prompt for analysis.
|
110 |
+
|
111 |
+
Returns:
|
112 |
+
str: The result of the analysis.
|
113 |
+
"""
|
114 |
dataset_info = f"""
|
115 |
Dataset Shape: {self.dataset.shape}
|
116 |
Columns: {', '.join(self.dataset.columns)}
|
|
|
132 |
|
133 |
@tool
|
134 |
def analyze_basic_stats(data: pd.DataFrame) -> str:
|
135 |
+
"""
|
136 |
+
Calculate and visualize basic statistical measures for numerical columns.
|
137 |
+
|
138 |
+
This function computes fundamental statistical metrics including mean, median,
|
139 |
+
standard deviation, skewness, and counts of missing values for all numerical
|
140 |
+
columns in the provided DataFrame. It also generates a bar chart visualizing
|
141 |
+
the mean, median, and standard deviation for each numerical feature.
|
142 |
+
|
143 |
+
Args:
|
144 |
+
data (pd.DataFrame): A pandas DataFrame containing the dataset to analyze.
|
145 |
+
The DataFrame should contain at least one numerical column
|
146 |
+
for meaningful analysis.
|
147 |
+
|
148 |
+
Returns:
|
149 |
+
str: A markdown-formatted string containing the statistics and the generated plot.
|
150 |
+
"""
|
151 |
if data is None:
|
152 |
data = tool.agent.dataset
|
153 |
|
|
|
185 |
|
186 |
@tool
|
187 |
def generate_correlation_matrix(data: pd.DataFrame) -> str:
|
188 |
+
"""
|
189 |
+
Generate an interactive correlation matrix using Plotly.
|
190 |
+
|
191 |
+
This function creates an interactive heatmap visualization showing the correlations between
|
192 |
+
all numerical columns in the dataset. Users can hover over cells to see correlation values
|
193 |
+
and interact with the plot (zoom, pan).
|
194 |
+
|
195 |
+
Args:
|
196 |
+
data (pd.DataFrame): A pandas DataFrame containing the dataset to analyze.
|
197 |
+
The DataFrame should contain at least two numerical columns
|
198 |
+
for correlation analysis.
|
199 |
+
|
200 |
+
Returns:
|
201 |
+
str: An HTML string representing the interactive correlation matrix plot.
|
202 |
+
"""
|
203 |
if data is None:
|
204 |
data = tool.agent.dataset
|
205 |
|
|
|
221 |
|
222 |
@tool
|
223 |
def analyze_categorical_columns(data: pd.DataFrame) -> str:
|
224 |
+
"""
|
225 |
+
Analyze categorical columns with visualizations.
|
226 |
+
|
227 |
+
This function examines categorical columns to identify unique values, top categories,
|
228 |
+
and missing value counts. It also generates bar charts for the top 5 categories in each
|
229 |
+
categorical feature.
|
230 |
+
|
231 |
+
Args:
|
232 |
+
data (pd.DataFrame): A pandas DataFrame containing the dataset to analyze.
|
233 |
+
The DataFrame should contain at least one categorical column
|
234 |
+
for meaningful analysis.
|
235 |
+
|
236 |
+
Returns:
|
237 |
+
str: A markdown-formatted string containing analysis results and embedded plots.
|
238 |
+
"""
|
239 |
if data is None:
|
240 |
data = tool.agent.dataset
|
241 |
|
|
|
276 |
|
277 |
@tool
|
278 |
def suggest_features(data: pd.DataFrame) -> str:
|
279 |
+
"""
|
280 |
+
Suggest potential feature engineering steps based on data characteristics.
|
281 |
+
|
282 |
+
This function analyzes the dataset's structure and statistical properties to
|
283 |
+
recommend possible feature engineering steps that could improve model performance.
|
284 |
+
|
285 |
+
Args:
|
286 |
+
data (pd.DataFrame): A pandas DataFrame containing the dataset to analyze.
|
287 |
+
The DataFrame can contain both numerical and categorical columns.
|
288 |
+
|
289 |
+
Returns:
|
290 |
+
str: A string containing suggestions for feature engineering based on
|
291 |
+
the characteristics of the input data.
|
292 |
+
"""
|
293 |
if data is None:
|
294 |
data = tool.agent.dataset
|
295 |
|
|
|
323 |
|
324 |
@tool
|
325 |
def predictive_analysis(data: pd.DataFrame, target: str) -> str:
|
326 |
+
"""
|
327 |
+
Perform predictive analytics by training a classification model.
|
328 |
+
|
329 |
+
This function builds a classification model using Random Forest, evaluates its performance,
|
330 |
+
and provides detailed metrics and visualizations such as the confusion matrix and ROC curve.
|
331 |
+
|
332 |
+
Args:
|
333 |
+
data (pd.DataFrame): A pandas DataFrame containing the dataset to analyze.
|
334 |
+
The DataFrame should contain the target variable for prediction.
|
335 |
+
target (str): The name of the target variable column in the dataset.
|
336 |
+
|
337 |
+
Returns:
|
338 |
+
str: A markdown-formatted string containing the classification report, confusion matrix,
|
339 |
+
ROC curve, AUC score, and a unique Model ID.
|
340 |
+
"""
|
341 |
if data is None:
|
342 |
data = tool.agent.dataset
|
343 |
|
|
|
432 |
# Report Exporting Function
|
433 |
# ------------------------------
|
434 |
def export_report(content: str, filename: str):
|
435 |
+
"""
|
436 |
+
Export the given content as a PDF report.
|
437 |
+
|
438 |
+
This function converts markdown content into a PDF file using pdfkit and provides
|
439 |
+
a download button for users to obtain the report.
|
440 |
+
|
441 |
+
Args:
|
442 |
+
content (str): The markdown content to be included in the PDF report.
|
443 |
+
filename (str): The desired name for the exported PDF file.
|
444 |
+
|
445 |
+
Returns:
|
446 |
+
None
|
447 |
+
"""
|
448 |
# Save content to a temporary HTML file
|
449 |
with tempfile.NamedTemporaryFile(delete=False, suffix='.html') as tmp_file:
|
450 |
tmp_file.write(content.encode('utf-8'))
|