hf SPACES TEST
Browse files- app.py +343 -0
- requirements.txt +11 -0
- run.py +39 -0
app.py
ADDED
@@ -0,0 +1,343 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
'''
|
2 |
+
Creator: Sudhir Arvind Deshmukh
|
3 |
+
Run command: streamlit run app.py
|
4 |
+
'''
|
5 |
+
import streamlit as st
|
6 |
+
import spacy
|
7 |
+
from spacy.tokens import Doc
|
8 |
+
from spacy.training.example import Example
|
9 |
+
import datetime
|
10 |
+
import os
|
11 |
+
import random
|
12 |
+
import pandas as pd
|
13 |
+
from sklearn.model_selection import train_test_split
|
14 |
+
import matplotlib.pyplot as plt
|
15 |
+
import datetime
|
16 |
+
from transformers import AutoTokenizer, T5ForConditionalGeneration
|
17 |
+
from spacy import displacy
|
18 |
+
|
19 |
+
## Load spaCy models from saved_models directory
|
20 |
+
|
21 |
+
# Get absolute path to the current script's directory
|
22 |
+
script_dir = os.path.dirname(os.path.abspath(__file__))
|
23 |
+
saved_models_dir = os.path.join(script_dir, "saved_models")
|
24 |
+
nlp_models = ["en_core_web_sm", "en_core_web_md", "en_core_web_lg"] + [os.path.join(saved_models_dir, str(model_name)) for model_name in os.listdir(saved_models_dir)]
|
25 |
+
|
26 |
+
# fuction to load the csv file and extract sentences and tags
|
27 |
+
def load_data_from_csv(file):
|
28 |
+
df = pd.read_csv(file, encoding="latin-1")
|
29 |
+
df = df.dropna()
|
30 |
+
#df.loc[:, "Sentence #"] = df["Sentence #"].fillna(method="ffill")
|
31 |
+
df.loc[:, "Sentence #"] = df["Sentence #"].ffill()
|
32 |
+
sentences = df.groupby("Sentence #")["Word"].apply(list).values
|
33 |
+
tags = df.groupby("Sentence #")["Tag"].apply(list).values
|
34 |
+
return sentences, tags
|
35 |
+
|
36 |
+
|
37 |
+
|
38 |
+
# Streamlit UI for Online Inference
|
39 |
+
def online_inference():
|
40 |
+
st.title("Online Inference")
|
41 |
+
|
42 |
+
selected_model = st.selectbox("Select base Model for finetunning", nlp_models)
|
43 |
+
|
44 |
+
# Load the selected spaCy model
|
45 |
+
# model_path = os.path.join(saved_models_dir, f"{selected_model}")
|
46 |
+
nlp = spacy.load(selected_model)
|
47 |
+
|
48 |
+
|
49 |
+
text_input = st.text_input("Enter Text for Inference")
|
50 |
+
|
51 |
+
if text_input:
|
52 |
+
doc = nlp(text_input)
|
53 |
+
|
54 |
+
# Filter out 'O' entities and get unique entity types
|
55 |
+
filtered_entities = [ent for ent in doc.ents if ent.label_ != 'O']
|
56 |
+
unique_entity_types = list(set(ent.label_ for ent in filtered_entities))
|
57 |
+
|
58 |
+
if filtered_entities:
|
59 |
+
# Define Google-themed colors for each entity type
|
60 |
+
color_dict = {
|
61 |
+
'B-geo': '#4285F4', # Blue
|
62 |
+
'B-gpe': '#EA4335', # Red
|
63 |
+
'B-per': '#FBBC05', # Yellow
|
64 |
+
'I-geo': '#0F9D58', # Green
|
65 |
+
'B-org': '#34A853', # Green
|
66 |
+
'I-org': '#FF9800', # Orange
|
67 |
+
'B-tim': '#AA66CC', # Purple
|
68 |
+
'B-art': '#FFC107', # Amber
|
69 |
+
'I-art': '#9C27B0', # Purple
|
70 |
+
'I-per': '#03A9F4', # Blue
|
71 |
+
'I-gpe': '#009688', # Teal
|
72 |
+
'I-tim': '#FF5722', # Deep Orange
|
73 |
+
'B-nat': '#7B1FA2', # Deep Purple
|
74 |
+
'B-eve': '#8BC34A', # Light Green
|
75 |
+
'I-eve': '#FDD835', # Yellow
|
76 |
+
'I-nat': '#616161' # Gray
|
77 |
+
}
|
78 |
+
|
79 |
+
# Render the visualization with custom colors
|
80 |
+
options = {"ents": unique_entity_types, "colors": color_dict}
|
81 |
+
html = spacy.displacy.render(doc, style="ent", options=options)
|
82 |
+
st.components.v1.html(html, height=400)
|
83 |
+
else:
|
84 |
+
st.write("No named entities found in the text.")
|
85 |
+
|
86 |
+
# Streamlit UI for Model Training
|
87 |
+
def model_training():
|
88 |
+
|
89 |
+
st.title("Model Training")
|
90 |
+
|
91 |
+
base_model = ["en_core_web_sm", "en_core_web_md", "en_core_web_lg"]
|
92 |
+
selected_model = st.selectbox("Select base Model to Train", base_model)
|
93 |
+
|
94 |
+
# Define hyperparameters
|
95 |
+
learning_rate = st.slider("Learning Rate", min_value=0.001, max_value=0.1, step=0.001, value=0.01)
|
96 |
+
n_iter = st.slider("Number of Iterations", min_value=1, max_value=10, value=2)
|
97 |
+
dropout = st.slider("Dropout", min_value=0.1, max_value=0.9, step=0.1, value=0.5)
|
98 |
+
|
99 |
+
uploaded_file = st.file_uploader("Upload Training Data (CSV)", type="csv")
|
100 |
+
|
101 |
+
model_name_uniq = st.text_input("Enter Model Name")
|
102 |
+
if st.button("Train & Evaluate Model"):
|
103 |
+
if uploaded_file is not None:
|
104 |
+
|
105 |
+
# Load training data from the uploaded CSV file
|
106 |
+
sentences, tags = load_data_from_csv(uploaded_file)
|
107 |
+
|
108 |
+
# Split data into training, validation, and test sets
|
109 |
+
train_sentences, test_sentences, train_tags, test_tags = train_test_split(sentences, tags, test_size=0.2, random_state=42)
|
110 |
+
train_sentences, val_sentences, train_tags, val_tags = train_test_split(train_sentences, train_tags, test_size=0.2, random_state=42)
|
111 |
+
|
112 |
+
print(f"Experimenting with model: {selected_model}")
|
113 |
+
|
114 |
+
|
115 |
+
# Load the pre-trained model
|
116 |
+
nlp = spacy.load(selected_model)
|
117 |
+
|
118 |
+
# Add or modify the NER component in the pipeline
|
119 |
+
if "ner" not in nlp.pipe_names:
|
120 |
+
ner = nlp.add_pipe("ner")
|
121 |
+
else:
|
122 |
+
ner = nlp.get_pipe("ner")
|
123 |
+
|
124 |
+
# Function to convert input format to spaCy format
|
125 |
+
def convert_to_spacy_format(sentences, tags):
|
126 |
+
examples = []
|
127 |
+
for sent, tag_list in zip(sentences, tags):
|
128 |
+
words = sent
|
129 |
+
spaces = [True] * len(words)
|
130 |
+
doc = Doc(nlp.vocab, words=words, spaces=spaces)
|
131 |
+
gold_entities = []
|
132 |
+
for token, tag in zip(doc, tag_list):
|
133 |
+
start = token.idx
|
134 |
+
end = start + len(token.text)
|
135 |
+
gold_entities.append((start, end, tag))
|
136 |
+
example = Example.from_dict(doc, {"entities": gold_entities})
|
137 |
+
examples.append(example)
|
138 |
+
return examples
|
139 |
+
|
140 |
+
# Add entity labels to the ner component
|
141 |
+
for label in set(tag for tag_list in tags for tag in tag_list):
|
142 |
+
ner.add_label(label)
|
143 |
+
|
144 |
+
# Create spaCy examples for training
|
145 |
+
train_examples = convert_to_spacy_format(train_sentences, train_tags)
|
146 |
+
val_examples = convert_to_spacy_format(val_sentences, val_tags)
|
147 |
+
|
148 |
+
|
149 |
+
# Lists to store learning curve data
|
150 |
+
train_losses = []
|
151 |
+
train_api_metrics = []
|
152 |
+
val_precisions = []
|
153 |
+
val_recalls = []
|
154 |
+
|
155 |
+
total_batches = len(train_examples) / 8
|
156 |
+
ner_metrics = []
|
157 |
+
# Train the NER model
|
158 |
+
for epoch in range(n_iter):
|
159 |
+
random.shuffle(train_examples)
|
160 |
+
st.write("this is iteration number:", epoch)
|
161 |
+
losses = {}
|
162 |
+
progress_bar = st.progress(0)
|
163 |
+
for batch_index, batch in enumerate(spacy.util.minibatch(train_examples, size=8), start=1):
|
164 |
+
nlp.update(batch, drop=dropout, losses=losses)
|
165 |
+
# Calculate progress percentage
|
166 |
+
progress_percentage = batch_index / (total_batches + 1)
|
167 |
+
progress_bar.progress(progress_percentage) # Display progress in Streamlit
|
168 |
+
train_losses.append(losses["ner"])
|
169 |
+
train_api_metrics.append(losses)
|
170 |
+
|
171 |
+
# Evaluate the model on the validation set
|
172 |
+
metrics = nlp.evaluate(val_examples)
|
173 |
+
val_precisions.append(metrics["ents_p"])
|
174 |
+
val_recalls.append(metrics["ents_r"])
|
175 |
+
|
176 |
+
# Append metrics to the ner_metrics list
|
177 |
+
ner_metrics.append(metrics)
|
178 |
+
print(val_precisions)
|
179 |
+
print(val_recalls)
|
180 |
+
current_time = datetime.datetime.now().strftime("%Y-%m-%d_%H-%M-%S")
|
181 |
+
save_model_name = f"{model_name_uniq}_ner_model_{current_time}"
|
182 |
+
# Plot learning curve
|
183 |
+
plt.figure(figsize=(12, 4))
|
184 |
+
plt.plot(range(n_iter), train_losses, label="Training Loss")
|
185 |
+
plt.xlabel("Epoch")
|
186 |
+
plt.ylabel("Loss")
|
187 |
+
plt.title(f"Learning Curve for Model: {save_model_name}")
|
188 |
+
plt.legend()
|
189 |
+
learning_curve_plot_path = f"images/learning_curve_{save_model_name}.png"
|
190 |
+
plt.savefig(learning_curve_plot_path)
|
191 |
+
st.image(learning_curve_plot_path)
|
192 |
+
|
193 |
+
# # Plot Precision-Recall curve (Not straight forward with spacy therefore lets do Brert implementation)
|
194 |
+
# plt.figure(figsize=(12, 4))
|
195 |
+
# plt.plot(val_recalls, val_precisions, label="Precision-Recall Curve")
|
196 |
+
# plt.xlabel("Recall")
|
197 |
+
# plt.ylabel("Precision")
|
198 |
+
# plt.title(f"Precision-Recall Curve for Model: {save_model_name}")
|
199 |
+
# plt.legend()
|
200 |
+
# pr_curve_plot_path = f"images/precision_recall_curve_{save_model_name}.png"
|
201 |
+
# plt.savefig(pr_curve_plot_path)
|
202 |
+
# st.image(pr_curve_plot_path)
|
203 |
+
|
204 |
+
# Save the trained model to disk with timestamp
|
205 |
+
|
206 |
+
nlp.to_disk(os.path.join(saved_models_dir, str(save_model_name)))
|
207 |
+
st.success(f"Trained model saved as: {save_model_name}")
|
208 |
+
|
209 |
+
# Print important NER performance metrics
|
210 |
+
ner_performance_metrics = ["ents_p", "ents_r", "ents_f",
|
211 |
+
#"ents_per_type"
|
212 |
+
]
|
213 |
+
# Print model performance metrics
|
214 |
+
st.write("---")
|
215 |
+
st.subheader("Evaluation Metrics on validation data (calculated during last epoch)")
|
216 |
+
for model_name, metrics in zip([selected_model], ner_metrics):
|
217 |
+
st.write(f"Model: {model_name}")
|
218 |
+
for metric_name in ner_performance_metrics:
|
219 |
+
metric_value = metrics.get(metric_name, 0.0)
|
220 |
+
st.write(f"{metric_name}: {metric_value}")
|
221 |
+
st.write("") # Add an empty line for spacing
|
222 |
+
st.write("---")
|
223 |
+
st.subheader("Performance Metrics on test data")
|
224 |
+
|
225 |
+
test_examples = convert_to_spacy_format(test_sentences, test_tags)
|
226 |
+
# Evaluate the model on the validation set
|
227 |
+
|
228 |
+
test_metrics = nlp.evaluate(test_examples)
|
229 |
+
# Print important NER performance metrics
|
230 |
+
# ner_performance_metrics = ["ents_p", "ents_r", "ents_f"]
|
231 |
+
# Print model performance metrics
|
232 |
+
|
233 |
+
for metric_name in ner_performance_metrics:
|
234 |
+
metric_value = test_metrics.get(metric_name, 0.0)
|
235 |
+
st.write(f"{metric_name}: {metric_value}")
|
236 |
+
|
237 |
+
st.write("---")
|
238 |
+
|
239 |
+
st.write(train_api_metrics)
|
240 |
+
st.write("training metric list of dicts")
|
241 |
+
st.write(ner_metrics)
|
242 |
+
st.write("training metric list of dicts")
|
243 |
+
st.write(test_metrics)
|
244 |
+
|
245 |
+
else:
|
246 |
+
st.warning("Please upload training data in CSV format.")
|
247 |
+
|
248 |
+
def gen_ai():
|
249 |
+
|
250 |
+
|
251 |
+
# Streamlit app layout
|
252 |
+
st.title("Few-Shot Named Entity Recognition with Flan")
|
253 |
+
|
254 |
+
# Load the Flan model
|
255 |
+
model_name = st.selectbox("Select Flan Model", ["google/flan-t5-base", "google/flan-t5-large", "google/flan-t5-xl", "google/flan-t5-xxl"])
|
256 |
+
model = T5ForConditionalGeneration.from_pretrained(model_name)
|
257 |
+
|
258 |
+
# Load a pre-trained tokenizer that's compatible with T5
|
259 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
260 |
+
st.write("---")
|
261 |
+
# User input for few-shot examples
|
262 |
+
st.subheader("Few-Shot Examples")
|
263 |
+
examples = []
|
264 |
+
num_examples = st.number_input("Number of Examples", min_value=1, value=2)
|
265 |
+
for _ in range(num_examples):
|
266 |
+
col1, col2 = st.columns([3, 1])
|
267 |
+
with col1:
|
268 |
+
example_text = st.text_input(f"Example {_+1} (Text)")
|
269 |
+
with col2:
|
270 |
+
example_label = st.text_input(f"Example {_+1} (Label)")
|
271 |
+
if example_text and example_label:
|
272 |
+
examples.append((example_text, example_label))
|
273 |
+
st.write("---")
|
274 |
+
# User input for query text
|
275 |
+
st.subheader("Query Text")
|
276 |
+
query = st.text_input("Enter Query Text")
|
277 |
+
|
278 |
+
# Detect Entities button
|
279 |
+
detect_button = st.button("Detect Entities")
|
280 |
+
|
281 |
+
# Generate named entities
|
282 |
+
if detect_button:
|
283 |
+
if not examples or not query:
|
284 |
+
st.warning('Need both examples and query as user input', icon="⚠️")
|
285 |
+
prompt = "\n".join([f"NER: {example[0]} Labels: {example[1]}" for example in examples])
|
286 |
+
prompt += f"\n{query} Labels:"
|
287 |
+
input_ids = tokenizer.encode(prompt, return_tensors="pt")
|
288 |
+
outputs = model.generate(input_ids, max_length=100, num_return_sequences=1)
|
289 |
+
generated_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
|
290 |
+
|
291 |
+
# Process the generated output for displacy
|
292 |
+
entities = generated_text.split("Labels:")
|
293 |
+
entities = [e.strip().split(":")[0].strip() for e in entities if e.strip()]
|
294 |
+
st.write("---")
|
295 |
+
# Display identified named entities
|
296 |
+
st.subheader("Identified Named Entities:")
|
297 |
+
|
298 |
+
doc = {"text": query, "ents": [{"start": query.find(entity), "end": query.find(entity) + len(entity), "label": "Custom Entity"} for entity in entities], "title": None}
|
299 |
+
html = displacy.render(doc, style="ent", manual=True, minify=True)
|
300 |
+
st.components.v1.html(html)
|
301 |
+
st.write("---")
|
302 |
+
st.write(doc)
|
303 |
+
|
304 |
+
|
305 |
+
def ensure_folders_exist(script_dir):
|
306 |
+
images_path = os.path.join(script_dir, "images")
|
307 |
+
saved_model_path = os.path.join(script_dir, "saved_models")
|
308 |
+
|
309 |
+
# Create the 'images' directory if it doesn't exist
|
310 |
+
if not os.path.exists(images_path):
|
311 |
+
os.makedirs(images_path)
|
312 |
+
|
313 |
+
# Create the 'saved_model' directory if it doesn't exist
|
314 |
+
if not os.path.exists(saved_model_path):
|
315 |
+
os.makedirs(saved_model_path)
|
316 |
+
|
317 |
+
def main():
|
318 |
+
|
319 |
+
script_dir = os.path.dirname(os.path.abspath(__file__))
|
320 |
+
|
321 |
+
# Ensure that required folders exist
|
322 |
+
ensure_folders_exist(script_dir)
|
323 |
+
|
324 |
+
# Streamlit App
|
325 |
+
st.set_page_config(page_title="NER Model Experimentation")
|
326 |
+
|
327 |
+
st.sidebar.title("Navigation")
|
328 |
+
page = st.sidebar.radio("Go to", ["Online Inference", "Model Training",
|
329 |
+
#"Evaluation Metrics",
|
330 |
+
"GEN AI"])
|
331 |
+
|
332 |
+
if page == "Online Inference":
|
333 |
+
online_inference()
|
334 |
+
elif page == "Model Training":
|
335 |
+
model_training()
|
336 |
+
elif page == "GEN AI":
|
337 |
+
gen_ai()
|
338 |
+
|
339 |
+
|
340 |
+
# call main fuction
|
341 |
+
if __name__=="__main__":
|
342 |
+
main()
|
343 |
+
|
requirements.txt
ADDED
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
numpy
|
2 |
+
pandas
|
3 |
+
streamlit
|
4 |
+
matplotlib
|
5 |
+
scikit_learn
|
6 |
+
torch
|
7 |
+
transformers
|
8 |
+
spacy==3.6.1
|
9 |
+
https://github.com/explosion/spacy-models/releases/download/en_core_web_sm-3.6.0/en_core_web_sm-3.6.0.tar.gz
|
10 |
+
https://github.com/explosion/spacy-models/releases/download/en_core_web_md-3.6.0/en_core_web_md-3.6.0.tar.gz
|
11 |
+
https://github.com/explosion/spacy-models/releases/download/en_core_web_lg-3.6.0/en_core_web_lg-3.6.0.tar.gz
|
run.py
ADDED
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
"""run the steamlit app through this python file"""
|
2 |
+
|
3 |
+
import os
|
4 |
+
import subprocess
|
5 |
+
import argparse
|
6 |
+
|
7 |
+
def ensure_folders_exist(script_dir):
|
8 |
+
images_path = os.path.join(script_dir, "images")
|
9 |
+
saved_model_path = os.path.join(script_dir, "saved_models")
|
10 |
+
|
11 |
+
# Create the 'images' directory if it doesn't exist
|
12 |
+
if not os.path.exists(images_path):
|
13 |
+
os.makedirs(images_path)
|
14 |
+
|
15 |
+
# Create the 'saved_model' directory if it doesn't exist
|
16 |
+
if not os.path.exists(saved_model_path):
|
17 |
+
os.makedirs(saved_model_path)
|
18 |
+
|
19 |
+
def run():
|
20 |
+
parser = argparse.ArgumentParser()
|
21 |
+
parser.add_argument("--port", type=int, default=8501,
|
22 |
+
help="Port number for the Streamlit app")
|
23 |
+
args = parser.parse_args()
|
24 |
+
|
25 |
+
# Get absolute path to the current script's directory
|
26 |
+
script_dir = os.path.dirname(os.path.abspath(__file__))
|
27 |
+
|
28 |
+
# Ensure that required folders exist
|
29 |
+
ensure_folders_exist(script_dir)
|
30 |
+
|
31 |
+
# Construct paths for app.py, images, and saved_model directories
|
32 |
+
app_path = os.path.join(script_dir, "app.py")
|
33 |
+
|
34 |
+
# Run the Streamlit app defined at app_path
|
35 |
+
cmd = ["python", "-m", "streamlit", "run", "--server.port", str(args.port), app_path]
|
36 |
+
subprocess.call(cmd)
|
37 |
+
|
38 |
+
if __name__ == "__main__":
|
39 |
+
run()
|