Spaces:
Sleeping
Sleeping
Yongdong
commited on
Commit
Β·
1ef829e
1
Parent(s):
792bd1c
Add DAG visualization functionality for robot task planning
Browse files- app.py +131 -52
- dag_visualizer.py +334 -0
- json_processor.py +46 -46
- requirements.txt +3 -0
- test_dag_integration.py +175 -0
app.py
CHANGED
@@ -3,6 +3,7 @@ import spaces # Import spaces module for ZeroGPU
|
|
3 |
from huggingface_hub import login
|
4 |
import os
|
5 |
from json_processor import JsonProcessor
|
|
|
6 |
import json
|
7 |
|
8 |
# 1) Read Secrets
|
@@ -213,6 +214,31 @@ def generate_response_gpu(prompt, max_tokens=512, selected_model=DEFAULT_MODEL):
|
|
213 |
except Exception as generation_error:
|
214 |
return f"β Generation Error: {str(generation_error)}"
|
215 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
216 |
def chat_interface(message, history, max_tokens, selected_model):
|
217 |
"""Chat interface - runs on CPU, calls GPU functions"""
|
218 |
if not message.strip():
|
@@ -255,7 +281,7 @@ with gr.Blocks(
|
|
255 |
- **βοΈ Dart-llm-model-3B**: Balanced performance and quality
|
256 |
- **π― Dart-llm-model-8B**: Best quality output, higher latency
|
257 |
|
258 |
-
**Capabilities**: Convert natural language robot commands into structured task sequences for excavators, dump trucks, and other construction robots.
|
259 |
|
260 |
**Models**:
|
261 |
- [YongdongWang/llama-3.2-1b-lora-qlora-dart-llm](https://huggingface.co/YongdongWang/llama-3.2-1b-lora-qlora-dart-llm) (Default)
|
@@ -265,59 +291,100 @@ with gr.Blocks(
|
|
265 |
β‘ **Using ZeroGPU**: This Space uses dynamic GPU allocation (Nvidia H200). First generation might take a bit longer.
|
266 |
""")
|
267 |
|
268 |
-
with gr.
|
269 |
-
with gr.
|
270 |
-
chatbot = gr.Chatbot(
|
271 |
-
label="Task Planning Results",
|
272 |
-
height=500,
|
273 |
-
show_label=True,
|
274 |
-
container=True,
|
275 |
-
bubble_full_width=False,
|
276 |
-
show_copy_button=True
|
277 |
-
)
|
278 |
-
|
279 |
-
msg = gr.Textbox(
|
280 |
-
label="Robot Command",
|
281 |
-
placeholder="Enter robot task command (e.g., 'Deploy Excavator 1 to Soil Area 1')...",
|
282 |
-
lines=2,
|
283 |
-
max_lines=5,
|
284 |
-
show_label=True,
|
285 |
-
container=True
|
286 |
-
)
|
287 |
-
|
288 |
with gr.Row():
|
289 |
-
|
290 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
291 |
|
292 |
-
with gr.
|
293 |
-
gr.
|
294 |
-
|
295 |
-
|
296 |
-
|
297 |
-
|
298 |
-
|
299 |
-
|
300 |
-
|
301 |
-
|
302 |
-
|
303 |
-
|
304 |
-
|
305 |
-
|
306 |
-
|
307 |
-
|
308 |
-
|
309 |
-
|
310 |
-
|
311 |
-
|
312 |
-
|
313 |
-
|
314 |
-
|
315 |
-
|
316 |
-
|
317 |
-
|
318 |
-
|
319 |
-
|
320 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
321 |
|
322 |
# Example conversations
|
323 |
gr.Examples(
|
@@ -350,6 +417,18 @@ with gr.Blocks(
|
|
350 |
lambda: ([], ""),
|
351 |
outputs=[chatbot, msg]
|
352 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
353 |
|
354 |
if __name__ == "__main__":
|
355 |
app.launch(
|
|
|
3 |
from huggingface_hub import login
|
4 |
import os
|
5 |
from json_processor import JsonProcessor
|
6 |
+
from dag_visualizer import DAGVisualizer
|
7 |
import json
|
8 |
|
9 |
# 1) Read Secrets
|
|
|
214 |
except Exception as generation_error:
|
215 |
return f"β Generation Error: {str(generation_error)}"
|
216 |
|
217 |
+
def create_dag_visualization(task_json_str):
|
218 |
+
"""Create DAG visualization from task JSON"""
|
219 |
+
try:
|
220 |
+
if not task_json_str.strip():
|
221 |
+
return None, "Please provide task JSON data"
|
222 |
+
|
223 |
+
# Parse JSON
|
224 |
+
task_data = json.loads(task_json_str)
|
225 |
+
|
226 |
+
# Create DAG visualizer
|
227 |
+
dag_visualizer = DAGVisualizer()
|
228 |
+
|
229 |
+
# Generate visualization
|
230 |
+
image_path = dag_visualizer.create_dag_visualization(task_data)
|
231 |
+
|
232 |
+
if image_path:
|
233 |
+
return image_path, "β
DAG visualization created successfully!"
|
234 |
+
else:
|
235 |
+
return None, "β Failed to create DAG visualization"
|
236 |
+
|
237 |
+
except json.JSONDecodeError as e:
|
238 |
+
return None, f"β JSON Parse Error: {str(e)}"
|
239 |
+
except Exception as e:
|
240 |
+
return None, f"β DAG Creation Error: {str(e)}"
|
241 |
+
|
242 |
def chat_interface(message, history, max_tokens, selected_model):
|
243 |
"""Chat interface - runs on CPU, calls GPU functions"""
|
244 |
if not message.strip():
|
|
|
281 |
- **βοΈ Dart-llm-model-3B**: Balanced performance and quality
|
282 |
- **π― Dart-llm-model-8B**: Best quality output, higher latency
|
283 |
|
284 |
+
**Capabilities**: Convert natural language robot commands into structured task sequences for excavators, dump trucks, and other construction robots. **Now with DAG Visualization!**
|
285 |
|
286 |
**Models**:
|
287 |
- [YongdongWang/llama-3.2-1b-lora-qlora-dart-llm](https://huggingface.co/YongdongWang/llama-3.2-1b-lora-qlora-dart-llm) (Default)
|
|
|
291 |
β‘ **Using ZeroGPU**: This Space uses dynamic GPU allocation (Nvidia H200). First generation might take a bit longer.
|
292 |
""")
|
293 |
|
294 |
+
with gr.Tabs():
|
295 |
+
with gr.Tab("π¬ Task Planning"):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
296 |
with gr.Row():
|
297 |
+
with gr.Column(scale=3):
|
298 |
+
chatbot = gr.Chatbot(
|
299 |
+
label="Task Planning Results",
|
300 |
+
height=500,
|
301 |
+
show_label=True,
|
302 |
+
container=True,
|
303 |
+
bubble_full_width=False,
|
304 |
+
show_copy_button=True
|
305 |
+
)
|
306 |
+
|
307 |
+
msg = gr.Textbox(
|
308 |
+
label="Robot Command",
|
309 |
+
placeholder="Enter robot task command (e.g., 'Deploy Excavator 1 to Soil Area 1')...",
|
310 |
+
lines=2,
|
311 |
+
max_lines=5,
|
312 |
+
show_label=True,
|
313 |
+
container=True
|
314 |
+
)
|
315 |
+
|
316 |
+
with gr.Row():
|
317 |
+
send_btn = gr.Button("π Generate Tasks", variant="primary", size="sm")
|
318 |
+
clear_btn = gr.Button("ποΈ Clear", variant="secondary", size="sm")
|
319 |
+
|
320 |
+
with gr.Column(scale=1):
|
321 |
+
gr.Markdown("### βοΈ Generation Settings")
|
322 |
+
|
323 |
+
model_selector = gr.Dropdown(
|
324 |
+
choices=[(config["name"], key) for key, config in MODEL_CONFIGS.items()],
|
325 |
+
value=DEFAULT_MODEL,
|
326 |
+
label="Model Size",
|
327 |
+
info="Select model size (1B = fastest, 8B = best quality)",
|
328 |
+
interactive=True
|
329 |
+
)
|
330 |
+
|
331 |
+
max_tokens = gr.Slider(
|
332 |
+
minimum=50,
|
333 |
+
maximum=5000,
|
334 |
+
value=512,
|
335 |
+
step=10,
|
336 |
+
label="Max Tokens",
|
337 |
+
info="Maximum number of tokens to generate"
|
338 |
+
)
|
339 |
+
|
340 |
+
gr.Markdown("""
|
341 |
+
### π Model Status
|
342 |
+
- **Hardware**: ZeroGPU (Dynamic Nvidia H200)
|
343 |
+
- **Status**: Ready
|
344 |
+
- **Note**: First generation allocates GPU resources
|
345 |
+
- **Dart-llm-model-1B**: Fastest inference (Default)
|
346 |
+
- **Dart-llm-model-3B**: Balanced speed/quality
|
347 |
+
- **Dart-llm-model-8B**: Best quality, slower
|
348 |
+
""")
|
349 |
|
350 |
+
with gr.Tab("π DAG Visualization"):
|
351 |
+
with gr.Row():
|
352 |
+
with gr.Column(scale=2):
|
353 |
+
json_input = gr.Textbox(
|
354 |
+
label="Task JSON Data",
|
355 |
+
placeholder="Paste the generated task JSON here to create a DAG visualization...",
|
356 |
+
lines=15,
|
357 |
+
max_lines=25,
|
358 |
+
show_label=True,
|
359 |
+
container=True
|
360 |
+
)
|
361 |
+
|
362 |
+
with gr.Row():
|
363 |
+
dag_btn = gr.Button("π¨ Generate DAG", variant="primary", size="sm")
|
364 |
+
dag_clear_btn = gr.Button("ποΈ Clear", variant="secondary", size="sm")
|
365 |
+
|
366 |
+
dag_status = gr.Textbox(
|
367 |
+
label="Status",
|
368 |
+
value="Ready to generate DAG visualization",
|
369 |
+
interactive=False,
|
370 |
+
show_label=True
|
371 |
+
)
|
372 |
+
|
373 |
+
with gr.Column(scale=3):
|
374 |
+
dag_output = gr.Image(
|
375 |
+
label="Task Dependency Graph",
|
376 |
+
show_label=True,
|
377 |
+
container=True,
|
378 |
+
height=600
|
379 |
+
)
|
380 |
+
|
381 |
+
gr.Markdown("""
|
382 |
+
### π DAG Features
|
383 |
+
- **Node Colors**: Red (Start), Orange (Intermediate), Purple (End)
|
384 |
+
- **Arrows**: Show task dependencies
|
385 |
+
- **Layout**: Hierarchical based on dependencies
|
386 |
+
- **Details**: Task info boxes with robots and objects
|
387 |
+
""")
|
388 |
|
389 |
# Example conversations
|
390 |
gr.Examples(
|
|
|
417 |
lambda: ([], ""),
|
418 |
outputs=[chatbot, msg]
|
419 |
)
|
420 |
+
|
421 |
+
# DAG visualization event handlers
|
422 |
+
dag_btn.click(
|
423 |
+
create_dag_visualization,
|
424 |
+
inputs=[json_input],
|
425 |
+
outputs=[dag_output, dag_status]
|
426 |
+
)
|
427 |
+
|
428 |
+
dag_clear_btn.click(
|
429 |
+
lambda: ("", None, "Ready to generate DAG visualization"),
|
430 |
+
outputs=[json_input, dag_output, dag_status]
|
431 |
+
)
|
432 |
|
433 |
if __name__ == "__main__":
|
434 |
app.launch(
|
dag_visualizer.py
ADDED
@@ -0,0 +1,334 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import matplotlib.pyplot as plt
|
2 |
+
import matplotlib
|
3 |
+
matplotlib.use('Agg') # Use non-interactive backend for server environments
|
4 |
+
import networkx as nx
|
5 |
+
import json
|
6 |
+
import numpy as np
|
7 |
+
from loguru import logger
|
8 |
+
import os
|
9 |
+
import tempfile
|
10 |
+
from datetime import datetime
|
11 |
+
|
12 |
+
class DAGVisualizer:
|
13 |
+
def __init__(self):
|
14 |
+
# Configure Matplotlib to use IEEE-style parameters
|
15 |
+
plt.rcParams.update({
|
16 |
+
'font.family': 'DejaVu Sans', # Use available font instead of Times New Roman
|
17 |
+
'font.size': 10,
|
18 |
+
'axes.linewidth': 1.2,
|
19 |
+
'axes.labelsize': 12,
|
20 |
+
'xtick.labelsize': 10,
|
21 |
+
'ytick.labelsize': 10,
|
22 |
+
'legend.fontsize': 10,
|
23 |
+
'figure.titlesize': 14
|
24 |
+
})
|
25 |
+
|
26 |
+
def create_dag_from_tasks(self, task_data):
|
27 |
+
"""
|
28 |
+
Create a directed graph from task data.
|
29 |
+
|
30 |
+
Args:
|
31 |
+
task_data: Dictionary containing tasks with structure like:
|
32 |
+
{
|
33 |
+
"tasks": [
|
34 |
+
{
|
35 |
+
"task": "task_name",
|
36 |
+
"instruction_function": {
|
37 |
+
"name": "function_name",
|
38 |
+
"robot_ids": ["robot1", "robot2"],
|
39 |
+
"dependencies": ["dependency_task"],
|
40 |
+
"object_keywords": ["object1", "object2"]
|
41 |
+
}
|
42 |
+
}
|
43 |
+
]
|
44 |
+
}
|
45 |
+
|
46 |
+
Returns:
|
47 |
+
NetworkX DiGraph object
|
48 |
+
"""
|
49 |
+
if not task_data or "tasks" not in task_data:
|
50 |
+
logger.warning("Invalid task data structure")
|
51 |
+
return None
|
52 |
+
|
53 |
+
# Create a directed graph
|
54 |
+
G = nx.DiGraph()
|
55 |
+
|
56 |
+
# Add nodes and store mapping from task name to ID
|
57 |
+
task_mapping = {}
|
58 |
+
for i, task in enumerate(task_data["tasks"]):
|
59 |
+
task_id = i + 1
|
60 |
+
task_name = task["task"]
|
61 |
+
task_mapping[task_name] = task_id
|
62 |
+
|
63 |
+
# Add node with attributes
|
64 |
+
G.add_node(task_id,
|
65 |
+
name=task_name,
|
66 |
+
function=task["instruction_function"]["name"],
|
67 |
+
robots=task["instruction_function"].get("robot_ids", []),
|
68 |
+
objects=task["instruction_function"].get("object_keywords", []))
|
69 |
+
|
70 |
+
# Add dependency edges
|
71 |
+
for i, task in enumerate(task_data["tasks"]):
|
72 |
+
task_id = i + 1
|
73 |
+
dependencies = task["instruction_function"]["dependencies"]
|
74 |
+
for dep in dependencies:
|
75 |
+
if dep in task_mapping:
|
76 |
+
dep_id = task_mapping[dep]
|
77 |
+
G.add_edge(dep_id, task_id)
|
78 |
+
|
79 |
+
return G
|
80 |
+
|
81 |
+
def calculate_layout(self, G):
|
82 |
+
"""
|
83 |
+
Calculate hierarchical layout for the graph based on dependencies.
|
84 |
+
"""
|
85 |
+
if not G:
|
86 |
+
return {}
|
87 |
+
|
88 |
+
# Calculate layers based on dependencies
|
89 |
+
layers = {}
|
90 |
+
|
91 |
+
def get_layer(node_id, visited=None):
|
92 |
+
if visited is None:
|
93 |
+
visited = set()
|
94 |
+
if node_id in visited:
|
95 |
+
return 0
|
96 |
+
visited.add(node_id)
|
97 |
+
|
98 |
+
predecessors = list(G.predecessors(node_id))
|
99 |
+
if not predecessors:
|
100 |
+
return 0
|
101 |
+
return max(get_layer(pred, visited.copy()) for pred in predecessors) + 1
|
102 |
+
|
103 |
+
for node in G.nodes():
|
104 |
+
layer = get_layer(node)
|
105 |
+
layers.setdefault(layer, []).append(node)
|
106 |
+
|
107 |
+
# Calculate positions by layer
|
108 |
+
pos = {}
|
109 |
+
layer_height = 3.0
|
110 |
+
node_width = 4.0
|
111 |
+
|
112 |
+
for layer_idx, nodes in layers.items():
|
113 |
+
y = layer_height * (len(layers) - 1 - layer_idx)
|
114 |
+
start_x = -(len(nodes) - 1) * node_width / 2
|
115 |
+
for i, node in enumerate(sorted(nodes)):
|
116 |
+
pos[node] = (start_x + i * node_width, y)
|
117 |
+
|
118 |
+
return pos
|
119 |
+
|
120 |
+
def create_dag_visualization(self, task_data, title="Robot Task Dependency Graph"):
|
121 |
+
"""
|
122 |
+
Create a DAG visualization from task data and return the image path.
|
123 |
+
|
124 |
+
Args:
|
125 |
+
task_data: Task data dictionary
|
126 |
+
title: Title for the graph
|
127 |
+
|
128 |
+
Returns:
|
129 |
+
str: Path to the generated image file
|
130 |
+
"""
|
131 |
+
try:
|
132 |
+
# Create graph
|
133 |
+
G = self.create_dag_from_tasks(task_data)
|
134 |
+
if not G or len(G.nodes()) == 0:
|
135 |
+
logger.warning("No tasks found or invalid graph structure")
|
136 |
+
return None
|
137 |
+
|
138 |
+
# Calculate layout
|
139 |
+
pos = self.calculate_layout(G)
|
140 |
+
|
141 |
+
# Create figure
|
142 |
+
fig, ax = plt.subplots(1, 1, figsize=(max(12, len(G.nodes()) * 2), 8))
|
143 |
+
|
144 |
+
# Draw edges with arrows
|
145 |
+
nx.draw_networkx_edges(G, pos,
|
146 |
+
edge_color='#2E86AB',
|
147 |
+
arrows=True,
|
148 |
+
arrowsize=20,
|
149 |
+
arrowstyle='->',
|
150 |
+
width=2,
|
151 |
+
alpha=0.8,
|
152 |
+
connectionstyle="arc3,rad=0.1")
|
153 |
+
|
154 |
+
# Color nodes based on their position in the graph
|
155 |
+
node_colors = []
|
156 |
+
for node in G.nodes():
|
157 |
+
if G.in_degree(node) == 0: # Start nodes
|
158 |
+
node_colors.append('#F24236')
|
159 |
+
elif G.out_degree(node) == 0: # End nodes
|
160 |
+
node_colors.append('#A23B72')
|
161 |
+
else: # Intermediate nodes
|
162 |
+
node_colors.append('#F18F01')
|
163 |
+
|
164 |
+
# Draw nodes
|
165 |
+
nx.draw_networkx_nodes(G, pos,
|
166 |
+
node_color=node_colors,
|
167 |
+
node_size=3500,
|
168 |
+
alpha=0.9,
|
169 |
+
edgecolors='black',
|
170 |
+
linewidths=2)
|
171 |
+
|
172 |
+
# Label nodes with task IDs
|
173 |
+
node_labels = {node: f"T{node}" for node in G.nodes()}
|
174 |
+
nx.draw_networkx_labels(G, pos, node_labels,
|
175 |
+
font_size=18,
|
176 |
+
font_weight='bold',
|
177 |
+
font_color='white')
|
178 |
+
|
179 |
+
# Add detailed info text boxes for each task
|
180 |
+
for i, node in enumerate(G.nodes()):
|
181 |
+
x, y = pos[node]
|
182 |
+
function_name = G.nodes[node]['function']
|
183 |
+
robots = G.nodes[node]['robots']
|
184 |
+
objects = G.nodes[node]['objects']
|
185 |
+
|
186 |
+
# Create info text content
|
187 |
+
info_text = f"Task {node}: {function_name.replace('_', ' ').title()}\n"
|
188 |
+
if robots:
|
189 |
+
robot_text = ", ".join([r.replace('robot_', '').replace('_', ' ').title() for r in robots])
|
190 |
+
info_text += f"Robots: {robot_text}\n"
|
191 |
+
if objects:
|
192 |
+
object_text = ", ".join(objects)
|
193 |
+
info_text += f"Objects: {object_text}"
|
194 |
+
|
195 |
+
# Calculate offset based on node position to avoid overlaps
|
196 |
+
offset_x = 2.2 if i % 2 == 0 else -2.2
|
197 |
+
offset_y = 0.5 if i % 4 < 2 else -0.5
|
198 |
+
|
199 |
+
# Choose alignment based on offset direction
|
200 |
+
h_align = 'left' if offset_x > 0 else 'right'
|
201 |
+
|
202 |
+
# Draw text box
|
203 |
+
bbox_props = dict(boxstyle="round,pad=0.4",
|
204 |
+
facecolor='white',
|
205 |
+
edgecolor='gray',
|
206 |
+
alpha=0.95,
|
207 |
+
linewidth=1)
|
208 |
+
|
209 |
+
ax.text(x + offset_x, y + offset_y, info_text,
|
210 |
+
bbox=bbox_props,
|
211 |
+
fontsize=12,
|
212 |
+
verticalalignment='center',
|
213 |
+
horizontalalignment=h_align,
|
214 |
+
weight='bold')
|
215 |
+
|
216 |
+
# Draw dashed connector line from node to text box
|
217 |
+
ax.plot([x, x + offset_x], [y, y + offset_y],
|
218 |
+
linestyle='--', color='gray', alpha=0.6, linewidth=1)
|
219 |
+
|
220 |
+
# Expand axis limits to fit everything
|
221 |
+
x_vals = [coord[0] for coord in pos.values()]
|
222 |
+
y_vals = [coord[1] for coord in pos.values()]
|
223 |
+
ax.set_xlim(min(x_vals) - 4.0, max(x_vals) + 4.0)
|
224 |
+
ax.set_ylim(min(y_vals) - 2.0, max(y_vals) + 2.0)
|
225 |
+
|
226 |
+
# Set overall figure properties
|
227 |
+
ax.set_title(title, fontsize=16, fontweight='bold', pad=20)
|
228 |
+
ax.set_aspect('equal')
|
229 |
+
ax.margins(0.2)
|
230 |
+
ax.axis('off')
|
231 |
+
|
232 |
+
# Add legend for node types - Hidden to avoid covering content
|
233 |
+
# legend_elements = [
|
234 |
+
# plt.Line2D([0], [0], marker='o', color='w', markerfacecolor='#F24236',
|
235 |
+
# markersize=10, label='Start Tasks', markeredgecolor='black'),
|
236 |
+
# plt.Line2D([0], [0], marker='o', color='w', markerfacecolor='#A23B72',
|
237 |
+
# markersize=10, label='End Tasks', markeredgecolor='black'),
|
238 |
+
# plt.Line2D([0], [0], marker='o', color='w', markerfacecolor='#F18F01',
|
239 |
+
# markersize=10, label='Intermediate Tasks', markeredgecolor='black'),
|
240 |
+
# plt.Line2D([0], [0], color='#2E86AB', linewidth=2, label='Dependencies')
|
241 |
+
# ]
|
242 |
+
# ax.legend(handles=legend_elements, loc='upper left', bbox_to_anchor=(1.05, 1.05))
|
243 |
+
|
244 |
+
# Adjust layout and save
|
245 |
+
plt.tight_layout()
|
246 |
+
|
247 |
+
# Create temporary file for saving the image
|
248 |
+
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
|
249 |
+
temp_dir = tempfile.gettempdir()
|
250 |
+
image_path = os.path.join(temp_dir, f'dag_visualization_{timestamp}.png')
|
251 |
+
|
252 |
+
plt.savefig(image_path, dpi=400, bbox_inches='tight',
|
253 |
+
pad_inches=0.1, facecolor='white', edgecolor='none')
|
254 |
+
plt.close(fig) # Close figure to free memory
|
255 |
+
|
256 |
+
logger.info(f"DAG visualization saved to: {image_path}")
|
257 |
+
return image_path
|
258 |
+
|
259 |
+
except Exception as e:
|
260 |
+
logger.error(f"Error creating DAG visualization: {e}")
|
261 |
+
return None
|
262 |
+
|
263 |
+
def create_simplified_dag_visualization(self, task_data, title="Robot Task Graph"):
|
264 |
+
"""
|
265 |
+
Create a simplified DAG visualization suitable for smaller displays.
|
266 |
+
|
267 |
+
Args:
|
268 |
+
task_data: Task data dictionary
|
269 |
+
title: Title for the graph
|
270 |
+
|
271 |
+
Returns:
|
272 |
+
str: Path to the generated image file
|
273 |
+
"""
|
274 |
+
try:
|
275 |
+
# Create graph
|
276 |
+
G = self.create_dag_from_tasks(task_data)
|
277 |
+
if not G or len(G.nodes()) == 0:
|
278 |
+
logger.warning("No tasks found or invalid graph structure")
|
279 |
+
return None
|
280 |
+
|
281 |
+
# Calculate layout
|
282 |
+
pos = self.calculate_layout(G)
|
283 |
+
|
284 |
+
# Create figure for simplified graph
|
285 |
+
fig, ax = plt.subplots(1, 1, figsize=(10, 6))
|
286 |
+
|
287 |
+
# Draw edges
|
288 |
+
nx.draw_networkx_edges(G, pos,
|
289 |
+
edge_color='black',
|
290 |
+
arrows=True,
|
291 |
+
arrowsize=15,
|
292 |
+
arrowstyle='->',
|
293 |
+
width=1.5)
|
294 |
+
|
295 |
+
# Draw nodes
|
296 |
+
nx.draw_networkx_nodes(G, pos,
|
297 |
+
node_color='lightblue',
|
298 |
+
node_size=3000,
|
299 |
+
edgecolors='black',
|
300 |
+
linewidths=1.5)
|
301 |
+
|
302 |
+
# Add node labels with simplified names
|
303 |
+
labels = {}
|
304 |
+
for node in G.nodes():
|
305 |
+
function_name = G.nodes[node]['function']
|
306 |
+
simplified_name = function_name.replace('_', ' ').title()
|
307 |
+
if len(simplified_name) > 15:
|
308 |
+
simplified_name = simplified_name[:12] + "..."
|
309 |
+
labels[node] = f"T{node}\n{simplified_name}"
|
310 |
+
|
311 |
+
nx.draw_networkx_labels(G, pos, labels,
|
312 |
+
font_size=11,
|
313 |
+
font_weight='bold')
|
314 |
+
|
315 |
+
ax.set_title(title, fontsize=14, fontweight='bold')
|
316 |
+
ax.axis('off')
|
317 |
+
|
318 |
+
# Adjust layout and save
|
319 |
+
plt.tight_layout()
|
320 |
+
|
321 |
+
# Create temporary file for saving the image
|
322 |
+
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
|
323 |
+
temp_dir = tempfile.gettempdir()
|
324 |
+
image_path = os.path.join(temp_dir, f'simple_dag_{timestamp}.png')
|
325 |
+
|
326 |
+
plt.savefig(image_path, dpi=400, bbox_inches='tight')
|
327 |
+
plt.close(fig) # Close figure to free memory
|
328 |
+
|
329 |
+
logger.info(f"Simplified DAG visualization saved to: {image_path}")
|
330 |
+
return image_path
|
331 |
+
|
332 |
+
except Exception as e:
|
333 |
+
logger.error(f"Error creating simplified DAG visualization: {e}")
|
334 |
+
return None
|
json_processor.py
CHANGED
@@ -1,46 +1,46 @@
|
|
1 |
-
import json
|
2 |
-
import re
|
3 |
-
import ast
|
4 |
-
from loguru import logger
|
5 |
-
|
6 |
-
class JsonProcessor:
|
7 |
-
def process_response(self, response):
|
8 |
-
try:
|
9 |
-
# Search for JSON string in the response
|
10 |
-
json_str_match = re.search(r'\{.*\}', response, re.DOTALL)
|
11 |
-
if json_str_match:
|
12 |
-
# Get the matched JSON string
|
13 |
-
json_str = json_str_match.group()
|
14 |
-
logger.debug(f"Full JSON string: {json_str}")
|
15 |
-
|
16 |
-
# Try to parse as Python literal first, then convert to JSON
|
17 |
-
try:
|
18 |
-
# First try to evaluate as Python literal
|
19 |
-
python_obj = ast.literal_eval(json_str)
|
20 |
-
# Convert to proper JSON
|
21 |
-
response_json = json.loads(json.dumps(python_obj))
|
22 |
-
except (ValueError, SyntaxError):
|
23 |
-
# Fall back to string replacement method
|
24 |
-
# Replace escape characters and remove trailing commas
|
25 |
-
json_str = json_str.replace("\\", "")
|
26 |
-
json_str = json_str.replace(r'\\_', '_')
|
27 |
-
json_str = re.sub(r',\s*}', '}', json_str)
|
28 |
-
json_str = re.sub(r',\s*\]', ']', json_str)
|
29 |
-
|
30 |
-
# Convert Python format to JSON format
|
31 |
-
json_str = json_str.replace("'", '"') # Single quotes to double quotes
|
32 |
-
json_str = json_str.replace('None', 'null') # Python None to JSON null
|
33 |
-
|
34 |
-
# Parse the JSON string
|
35 |
-
response_json = json.loads(json_str)
|
36 |
-
return response_json
|
37 |
-
else:
|
38 |
-
logger.error("No JSON string match found in response.")
|
39 |
-
return None
|
40 |
-
|
41 |
-
except json.JSONDecodeError as e:
|
42 |
-
logger.error(f"JSONDecodeError: {e}")
|
43 |
-
except Exception as e:
|
44 |
-
logger.error(f"Unexpected error: {e}")
|
45 |
-
|
46 |
-
return None
|
|
|
1 |
+
import json
|
2 |
+
import re
|
3 |
+
import ast
|
4 |
+
from loguru import logger
|
5 |
+
|
6 |
+
class JsonProcessor:
|
7 |
+
def process_response(self, response):
|
8 |
+
try:
|
9 |
+
# Search for JSON string in the response
|
10 |
+
json_str_match = re.search(r'\{.*\}', response, re.DOTALL)
|
11 |
+
if json_str_match:
|
12 |
+
# Get the matched JSON string
|
13 |
+
json_str = json_str_match.group()
|
14 |
+
logger.debug(f"Full JSON string: {json_str}")
|
15 |
+
|
16 |
+
# Try to parse as Python literal first, then convert to JSON
|
17 |
+
try:
|
18 |
+
# First try to evaluate as Python literal
|
19 |
+
python_obj = ast.literal_eval(json_str)
|
20 |
+
# Convert to proper JSON
|
21 |
+
response_json = json.loads(json.dumps(python_obj))
|
22 |
+
except (ValueError, SyntaxError):
|
23 |
+
# Fall back to string replacement method
|
24 |
+
# Replace escape characters and remove trailing commas
|
25 |
+
json_str = json_str.replace("\\", "")
|
26 |
+
json_str = json_str.replace(r'\\_', '_')
|
27 |
+
json_str = re.sub(r',\s*}', '}', json_str)
|
28 |
+
json_str = re.sub(r',\s*\]', ']', json_str)
|
29 |
+
|
30 |
+
# Convert Python format to JSON format
|
31 |
+
json_str = json_str.replace("'", '"') # Single quotes to double quotes
|
32 |
+
json_str = json_str.replace('None', 'null') # Python None to JSON null
|
33 |
+
|
34 |
+
# Parse the JSON string
|
35 |
+
response_json = json.loads(json_str)
|
36 |
+
return response_json
|
37 |
+
else:
|
38 |
+
logger.error("No JSON string match found in response.")
|
39 |
+
return None
|
40 |
+
|
41 |
+
except json.JSONDecodeError as e:
|
42 |
+
logger.error(f"JSONDecodeError: {e}")
|
43 |
+
except Exception as e:
|
44 |
+
logger.error(f"Unexpected error: {e}")
|
45 |
+
|
46 |
+
return None
|
requirements.txt
CHANGED
@@ -10,3 +10,6 @@ sentencepiece
|
|
10 |
protobuf
|
11 |
spaces
|
12 |
loguru
|
|
|
|
|
|
|
|
10 |
protobuf
|
11 |
spaces
|
12 |
loguru
|
13 |
+
matplotlib
|
14 |
+
networkx
|
15 |
+
numpy
|
test_dag_integration.py
ADDED
@@ -0,0 +1,175 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
#!/usr/bin/env python3
|
2 |
+
"""
|
3 |
+
Test script for DAG integration in DART-LLM-Multi-Model
|
4 |
+
"""
|
5 |
+
|
6 |
+
from dag_visualizer import DAGVisualizer
|
7 |
+
from json_processor import JsonProcessor
|
8 |
+
import json
|
9 |
+
|
10 |
+
def test_dag_integration():
|
11 |
+
"""Test the DAG integration with sample task data"""
|
12 |
+
print("Testing DAG integration...")
|
13 |
+
|
14 |
+
# Sample response with task data (similar to what the model might generate)
|
15 |
+
sample_response = """
|
16 |
+
Based on your command, here are the robot tasks:
|
17 |
+
|
18 |
+
{
|
19 |
+
"tasks": [
|
20 |
+
{
|
21 |
+
"task": "move_excavator_to_soil_area",
|
22 |
+
"instruction_function": {
|
23 |
+
"name": "move_to_position",
|
24 |
+
"robot_ids": ["robot_excavator_01"],
|
25 |
+
"dependencies": [],
|
26 |
+
"object_keywords": ["soil_area_1"]
|
27 |
+
}
|
28 |
+
},
|
29 |
+
{
|
30 |
+
"task": "excavate_soil",
|
31 |
+
"instruction_function": {
|
32 |
+
"name": "excavate_material",
|
33 |
+
"robot_ids": ["robot_excavator_01"],
|
34 |
+
"dependencies": ["move_excavator_to_soil_area"],
|
35 |
+
"object_keywords": ["soil"]
|
36 |
+
}
|
37 |
+
},
|
38 |
+
{
|
39 |
+
"task": "move_dump_truck",
|
40 |
+
"instruction_function": {
|
41 |
+
"name": "move_to_position",
|
42 |
+
"robot_ids": ["robot_dump_truck_01"],
|
43 |
+
"dependencies": [],
|
44 |
+
"object_keywords": ["soil_area_1"]
|
45 |
+
}
|
46 |
+
},
|
47 |
+
{
|
48 |
+
"task": "load_soil_to_truck",
|
49 |
+
"instruction_function": {
|
50 |
+
"name": "transfer_material",
|
51 |
+
"robot_ids": ["robot_excavator_01", "robot_dump_truck_01"],
|
52 |
+
"dependencies": ["excavate_soil", "move_dump_truck"],
|
53 |
+
"object_keywords": ["soil"]
|
54 |
+
}
|
55 |
+
},
|
56 |
+
{
|
57 |
+
"task": "transport_to_dump_site",
|
58 |
+
"instruction_function": {
|
59 |
+
"name": "move_to_position",
|
60 |
+
"robot_ids": ["robot_dump_truck_01"],
|
61 |
+
"dependencies": ["load_soil_to_truck"],
|
62 |
+
"object_keywords": ["dump_site"]
|
63 |
+
}
|
64 |
+
}
|
65 |
+
]
|
66 |
+
}
|
67 |
+
"""
|
68 |
+
|
69 |
+
# Test JSON processing
|
70 |
+
processor = JsonProcessor()
|
71 |
+
print("1. Testing JSON processing...")
|
72 |
+
processed_json = processor.process_response(sample_response)
|
73 |
+
|
74 |
+
if processed_json:
|
75 |
+
print("β JSON processing successful")
|
76 |
+
print(f" Found {len(processed_json['tasks'])} tasks")
|
77 |
+
else:
|
78 |
+
print("β JSON processing failed")
|
79 |
+
return False
|
80 |
+
|
81 |
+
# Test DAG visualization
|
82 |
+
print("2. Testing DAG visualization...")
|
83 |
+
visualizer = DAGVisualizer()
|
84 |
+
|
85 |
+
try:
|
86 |
+
dag_image_path = visualizer.create_dag_visualization(
|
87 |
+
processed_json,
|
88 |
+
title="Test Robot Task Dependency Graph"
|
89 |
+
)
|
90 |
+
|
91 |
+
if dag_image_path:
|
92 |
+
print(f"β DAG visualization created: {dag_image_path}")
|
93 |
+
return True
|
94 |
+
else:
|
95 |
+
print("β DAG visualization failed")
|
96 |
+
return False
|
97 |
+
|
98 |
+
except Exception as e:
|
99 |
+
print(f"β DAG visualization error: {e}")
|
100 |
+
return False
|
101 |
+
|
102 |
+
def test_simplified_dag():
|
103 |
+
"""Test simplified DAG visualization"""
|
104 |
+
print("\n3. Testing simplified DAG...")
|
105 |
+
|
106 |
+
simple_task_data = {
|
107 |
+
"tasks": [
|
108 |
+
{
|
109 |
+
"task": "move_robot",
|
110 |
+
"instruction_function": {
|
111 |
+
"name": "move_to_position",
|
112 |
+
"robot_ids": ["robot_01"],
|
113 |
+
"dependencies": [],
|
114 |
+
"object_keywords": ["target_area"]
|
115 |
+
}
|
116 |
+
},
|
117 |
+
{
|
118 |
+
"task": "perform_operation",
|
119 |
+
"instruction_function": {
|
120 |
+
"name": "excavate",
|
121 |
+
"robot_ids": ["robot_01"],
|
122 |
+
"dependencies": ["move_robot"],
|
123 |
+
"object_keywords": ["soil"]
|
124 |
+
}
|
125 |
+
}
|
126 |
+
]
|
127 |
+
}
|
128 |
+
|
129 |
+
visualizer = DAGVisualizer()
|
130 |
+
|
131 |
+
try:
|
132 |
+
dag_image_path = visualizer.create_simplified_dag_visualization(
|
133 |
+
simple_task_data,
|
134 |
+
title="Simplified Test DAG"
|
135 |
+
)
|
136 |
+
|
137 |
+
if dag_image_path:
|
138 |
+
print(f"β Simplified DAG visualization created: {dag_image_path}")
|
139 |
+
return True
|
140 |
+
else:
|
141 |
+
print("β Simplified DAG visualization failed")
|
142 |
+
return False
|
143 |
+
|
144 |
+
except Exception as e:
|
145 |
+
print(f"β Simplified DAG visualization error: {e}")
|
146 |
+
return False
|
147 |
+
|
148 |
+
def main():
|
149 |
+
"""Run all tests"""
|
150 |
+
print("=" * 60)
|
151 |
+
print("DART-LLM-Multi-Model DAG Integration Test")
|
152 |
+
print("=" * 60)
|
153 |
+
|
154 |
+
success_count = 0
|
155 |
+
total_tests = 2
|
156 |
+
|
157 |
+
if test_dag_integration():
|
158 |
+
success_count += 1
|
159 |
+
|
160 |
+
if test_simplified_dag():
|
161 |
+
success_count += 1
|
162 |
+
|
163 |
+
print("\n" + "=" * 60)
|
164 |
+
print(f"Test Results: {success_count}/{total_tests} passed")
|
165 |
+
|
166 |
+
if success_count == total_tests:
|
167 |
+
print("π All DAG integration tests passed!")
|
168 |
+
return True
|
169 |
+
else:
|
170 |
+
print("β Some tests failed!")
|
171 |
+
return False
|
172 |
+
|
173 |
+
if __name__ == "__main__":
|
174 |
+
success = main()
|
175 |
+
exit(0 if success else 1)
|