levisstrauss commited on
Commit
ea1c770
·
1 Parent(s): 92bdf98

Initial deployment of AI Project Planner

Browse files
Files changed (5) hide show
  1. .gitignore +1 -0
  2. README.md +120 -8
  3. app.py +1143 -0
  4. base_agents.py +525 -0
  5. requirements.txt +14 -0
.gitignore ADDED
@@ -0,0 +1 @@
 
 
1
+ .idea/
README.md CHANGED
@@ -1,14 +1,126 @@
1
  ---
2
- title: Ai Project Planner
3
- emoji: 📊
4
- colorFrom: pink
5
- colorTo: indigo
6
  sdk: gradio
7
- sdk_version: 6.2.0
8
  app_file: app.py
9
- pinned: false
10
  license: mit
11
- short_description: An AI planner project planner
12
  ---
13
 
14
- Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
+ title: AI Project Planner
3
+ emoji: 🚀
4
+ colorFrom: purple
5
+ colorTo: blue
6
  sdk: gradio
7
+ sdk_version: 4.0.0
8
  app_file: app.py
9
+ pinned: true
10
  license: mit
11
+ short_description: Transform product specs into comprehensive project plans using AI agents
12
  ---
13
 
14
+ # 🚀 AI Project Planner
15
+
16
+ **Transform product specifications into comprehensive project plans using orchestrated AI agents.**
17
+
18
+ [![Hugging Face Spaces](https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Spaces-blue)](https://huggingface.co/spaces)
19
+ [![Python 3.8+](https://img.shields.io/badge/python-3.8+-blue.svg)](https://www.python.org/downloads/)
20
+ [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
21
+
22
+ ## What This Does
23
+
24
+ Upload any product specification and watch as specialized AI agents collaborate to generate:
25
+
26
+ - **👤 User Stories** - Persona-based requirements in "As a... I want... So that..." format
27
+ - **📦 Product Features** - Logically grouped capabilities with descriptions and user benefits
28
+ - **🔧 Engineering Tasks** - Detailed development work breakdown with dependencies and estimates
29
+
30
+ ## How It Works
31
+
32
+ ```
33
+ Product Spec → Action Planner → Routing Agent → Specialized Teams → Project Plan
34
+ │
35
+ ┌─────────────────┼─────────────────┐
36
+ ↓ ↓ ↓
37
+ Product Manager Program Manager Dev Engineer
38
+ (User Stories) (Features) (Tasks)
39
+ │ │ │
40
+ └────────────────►│◄────────────────┘
41
+ Evaluation Agents
42
+ (Quality Assurance Loop)
43
+ ```
44
+
45
+ ### The Agent Architecture
46
+
47
+ | Agent | Role | Output |
48
+ |-------|------|--------|
49
+ | **Action Planning Agent** | Breaks down high-level goals into workflow steps | Task sequence |
50
+ | **Routing Agent** | Directs each step to the appropriate specialist using semantic similarity | Agent selection |
51
+ | **Product Manager Team** | Knowledge Agent + Evaluation Agent for user stories | Validated stories |
52
+ | **Program Manager Team** | Knowledge Agent + Evaluation Agent for features | Validated features |
53
+ | **Dev Engineer Team** | Knowledge Agent + Evaluation Agent for tasks | Validated tasks |
54
+
55
+ ## Quick Start
56
+
57
+ 1. **Enter your OpenAI API key** (required for AI agents to function)
58
+ 2. **Upload or paste** a product specification (or use an example)
59
+ 3. **Click Generate** and watch the workflow execute
60
+ 4. **Download** your project plan in Markdown or JSON format
61
+
62
+ ## Example Output
63
+
64
+ ### User Story
65
+ ```
66
+ As a Customer Support Representative, I want the Email Router system
67
+ to automate responses to routine inquiries so that I can focus on
68
+ addressing complex customer issues efficiently.
69
+ ```
70
+
71
+ ### Product Feature
72
+ ```
73
+ Feature Name: Intelligent Email Classification
74
+ Description: NLP-powered analysis for categorizing incoming emails
75
+ Key Functionality: Content analysis, intent detection, priority assignment
76
+ User Benefit: Reduced manual triage time by 70%
77
+ ```
78
+
79
+ ### Engineering Task
80
+ ```
81
+ Task ID: EMAIL-001
82
+ Task Title: Implement NLP classification engine
83
+ Related User Story: US-001
84
+ Description: Develop ML pipeline for email content analysis
85
+ Acceptance Criteria: Classification accuracy >95%
86
+ Estimated Effort: 3 weeks
87
+ Dependencies: Email preprocessing module
88
+ ```
89
+
90
+ ## Technical Details
91
+
92
+ - **Model**: GPT-3.5-turbo for text generation
93
+ - **Embeddings**: text-embedding-3-large for semantic routing
94
+ - **Framework**: Gradio for the web interface
95
+ - **Quality Control**: Iterative evaluation loops with up to 5 refinement cycles
96
+
97
+ ## Use Cases
98
+
99
+ - **Product Managers**: Quickly generate user stories from feature ideas
100
+ - **Technical Program Managers**: Create comprehensive project plans
101
+ - **Engineering Leads**: Break down requirements into actionable tasks
102
+ - **Startups**: Transform product vision into development roadmaps
103
+
104
+ ## Privacy & Security
105
+
106
+ - Your API key is **never stored** - it's only used for the current session
107
+ - Product specifications are **not retained** after processing
108
+ - All processing happens in real-time with no data persistence
109
+
110
+ ## Built With
111
+
112
+ - [Gradio](https://gradio.app/) - Web interface framework
113
+ - [OpenAI API](https://openai.com/) - AI capabilities
114
+ - [NumPy](https://numpy.org/) - Vector operations for semantic routing
115
+
116
+ ## Author
117
+
118
+ Built as part of the Udacity AI Programming curriculum, extended with production deployment capabilities.
119
+
120
+ ## License
121
+
122
+ MIT License - feel free to use, modify, and distribute.
123
+
124
+ ---
125
+
126
+ **⭐ If you find this useful, please star the repository!**
app.py ADDED
@@ -0,0 +1,1143 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ """
2
+ AI Project Planner - Agentic Workflow for Technical Project Management
3
+
4
+ A professional Gradio application that transforms product specifications into
5
+ comprehensive project plans using orchestrated AI agents.
6
+
7
+ Author: AI Workflow Architect
8
+ License: MIT
9
+ """
10
+
11
+ import gradio as gr
12
+ import json
13
+ import time
14
+ from datetime import datetime
15
+ from typing import Generator, Tuple, Dict, Any, List
16
+ import os
17
+
18
+ # Import the agent library
19
+ from base_agents import (
20
+ ActionPlanningAgent,
21
+ KnowledgeAugmentedPromptAgent,
22
+ EvaluationAgent,
23
+ RoutingAgent
24
+ )
25
+
26
+ # ============================================================================
27
+ # CUSTOM CSS FOR PROFESSIONAL STYLING
28
+ # ============================================================================
29
+
30
+ CUSTOM_CSS = """
31
+ /* Import Google Fonts */
32
+ @import url('https://fonts.googleapis.com/css2?family=Space+Grotesk:wght@300;400;500;600;700&family=JetBrains+Mono:wght@400;500&display=swap');
33
+
34
+ /* Root Variables */
35
+ :root {
36
+ --primary-gradient: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
37
+ --secondary-gradient: linear-gradient(135deg, #f093fb 0%, #f5576c 100%);
38
+ --success-gradient: linear-gradient(135deg, #4facfe 0%, #00f2fe 100%);
39
+ --dark-bg: #0f0f1a;
40
+ --card-bg: #1a1a2e;
41
+ --card-border: #2d2d44;
42
+ --text-primary: #ffffff;
43
+ --text-secondary: #a0a0b0;
44
+ --accent-purple: #667eea;
45
+ --accent-pink: #f093fb;
46
+ --accent-cyan: #00f2fe;
47
+ }
48
+
49
+ /* Global Styles */
50
+ .gradio-container {
51
+ font-family: 'Space Grotesk', sans-serif !important;
52
+ background: var(--dark-bg) !important;
53
+ max-width: 1400px !important;
54
+ }
55
+
56
+ /* Header Styling */
57
+ .header-container {
58
+ background: linear-gradient(135deg, rgba(102, 126, 234, 0.1) 0%, rgba(118, 75, 162, 0.1) 100%);
59
+ border: 1px solid var(--card-border);
60
+ border-radius: 20px;
61
+ padding: 2rem;
62
+ margin-bottom: 1.5rem;
63
+ position: relative;
64
+ overflow: hidden;
65
+ }
66
+
67
+ .header-container::before {
68
+ content: '';
69
+ position: absolute;
70
+ top: 0;
71
+ left: 0;
72
+ right: 0;
73
+ height: 3px;
74
+ background: var(--primary-gradient);
75
+ }
76
+
77
+ .header-title {
78
+ font-size: 2.5rem;
79
+ font-weight: 700;
80
+ background: var(--primary-gradient);
81
+ -webkit-background-clip: text;
82
+ -webkit-text-fill-color: transparent;
83
+ background-clip: text;
84
+ margin-bottom: 0.5rem;
85
+ }
86
+
87
+ .header-subtitle {
88
+ color: var(--text-secondary);
89
+ font-size: 1.1rem;
90
+ font-weight: 400;
91
+ }
92
+
93
+ /* Card Styling */
94
+ .input-card, .output-card, .workflow-card {
95
+ background: var(--card-bg) !important;
96
+ border: 1px solid var(--card-border) !important;
97
+ border-radius: 16px !important;
98
+ padding: 1.5rem !important;
99
+ }
100
+
101
+ /* Workflow Progress */
102
+ .workflow-step {
103
+ display: flex;
104
+ align-items: center;
105
+ padding: 1rem;
106
+ margin: 0.5rem 0;
107
+ border-radius: 12px;
108
+ background: rgba(255, 255, 255, 0.02);
109
+ border: 1px solid transparent;
110
+ transition: all 0.3s ease;
111
+ }
112
+
113
+ .workflow-step.active {
114
+ background: rgba(102, 126, 234, 0.1);
115
+ border-color: var(--accent-purple);
116
+ }
117
+
118
+ .workflow-step.completed {
119
+ background: rgba(0, 242, 254, 0.1);
120
+ border-color: var(--accent-cyan);
121
+ }
122
+
123
+ .step-icon {
124
+ width: 40px;
125
+ height: 40px;
126
+ border-radius: 50%;
127
+ display: flex;
128
+ align-items: center;
129
+ justify-content: center;
130
+ margin-right: 1rem;
131
+ font-size: 1.2rem;
132
+ }
133
+
134
+ .step-icon.waiting {
135
+ background: rgba(255, 255, 255, 0.1);
136
+ }
137
+
138
+ .step-icon.active {
139
+ background: var(--primary-gradient);
140
+ animation: pulse 2s infinite;
141
+ }
142
+
143
+ .step-icon.completed {
144
+ background: var(--success-gradient);
145
+ }
146
+
147
+ @keyframes pulse {
148
+ 0%, 100% { transform: scale(1); opacity: 1; }
149
+ 50% { transform: scale(1.05); opacity: 0.8; }
150
+ }
151
+
152
+ /* Button Styling */
153
+ .primary-btn {
154
+ background: var(--primary-gradient) !important;
155
+ border: none !important;
156
+ border-radius: 12px !important;
157
+ padding: 1rem 2rem !important;
158
+ font-weight: 600 !important;
159
+ font-size: 1rem !important;
160
+ transition: all 0.3s ease !important;
161
+ box-shadow: 0 4px 15px rgba(102, 126, 234, 0.3) !important;
162
+ }
163
+
164
+ .primary-btn:hover {
165
+ transform: translateY(-2px) !important;
166
+ box-shadow: 0 6px 20px rgba(102, 126, 234, 0.4) !important;
167
+ }
168
+
169
+ .secondary-btn {
170
+ background: transparent !important;
171
+ border: 1px solid var(--card-border) !important;
172
+ border-radius: 12px !important;
173
+ color: var(--text-secondary) !important;
174
+ transition: all 0.3s ease !important;
175
+ }
176
+
177
+ .secondary-btn:hover {
178
+ border-color: var(--accent-purple) !important;
179
+ color: var(--accent-purple) !important;
180
+ }
181
+
182
+ /* Tab Styling */
183
+ .tabs {
184
+ border: none !important;
185
+ }
186
+
187
+ .tab-nav {
188
+ background: transparent !important;
189
+ border-bottom: 1px solid var(--card-border) !important;
190
+ gap: 0.5rem !important;
191
+ }
192
+
193
+ .tab-nav button {
194
+ background: transparent !important;
195
+ border: none !important;
196
+ border-radius: 8px 8px 0 0 !important;
197
+ color: var(--text-secondary) !important;
198
+ padding: 0.75rem 1.5rem !important;
199
+ font-weight: 500 !important;
200
+ transition: all 0.3s ease !important;
201
+ }
202
+
203
+ .tab-nav button.selected {
204
+ background: rgba(102, 126, 234, 0.2) !important;
205
+ color: var(--accent-purple) !important;
206
+ border-bottom: 2px solid var(--accent-purple) !important;
207
+ }
208
+
209
+ /* Textbox Styling */
210
+ .textbox textarea {
211
+ background: rgba(0, 0, 0, 0.3) !important;
212
+ border: 1px solid var(--card-border) !important;
213
+ border-radius: 12px !important;
214
+ color: var(--text-primary) !important;
215
+ font-family: 'JetBrains Mono', monospace !important;
216
+ font-size: 0.9rem !important;
217
+ padding: 1rem !important;
218
+ }
219
+
220
+ .textbox textarea:focus {
221
+ border-color: var(--accent-purple) !important;
222
+ box-shadow: 0 0 0 2px rgba(102, 126, 234, 0.2) !important;
223
+ }
224
+
225
+ /* Markdown Styling */
226
+ .markdown-text {
227
+ color: var(--text-primary) !important;
228
+ }
229
+
230
+ .markdown-text h1, .markdown-text h2, .markdown-text h3 {
231
+ color: var(--text-primary) !important;
232
+ font-weight: 600 !important;
233
+ }
234
+
235
+ .markdown-text code {
236
+ background: rgba(102, 126, 234, 0.2) !important;
237
+ color: var(--accent-cyan) !important;
238
+ padding: 0.2rem 0.5rem !important;
239
+ border-radius: 4px !important;
240
+ font-family: 'JetBrains Mono', monospace !important;
241
+ }
242
+
243
+ /* Stats Card */
244
+ .stats-container {
245
+ display: grid;
246
+ grid-template-columns: repeat(4, 1fr);
247
+ gap: 1rem;
248
+ margin-top: 1rem;
249
+ }
250
+
251
+ .stat-card {
252
+ background: rgba(255, 255, 255, 0.02);
253
+ border: 1px solid var(--card-border);
254
+ border-radius: 12px;
255
+ padding: 1rem;
256
+ text-align: center;
257
+ }
258
+
259
+ .stat-value {
260
+ font-size: 1.5rem;
261
+ font-weight: 700;
262
+ background: var(--primary-gradient);
263
+ -webkit-background-clip: text;
264
+ -webkit-text-fill-color: transparent;
265
+ }
266
+
267
+ .stat-label {
268
+ font-size: 0.85rem;
269
+ color: var(--text-secondary);
270
+ margin-top: 0.25rem;
271
+ }
272
+
273
+ /* Agent Badge */
274
+ .agent-badge {
275
+ display: inline-flex;
276
+ align-items: center;
277
+ padding: 0.25rem 0.75rem;
278
+ border-radius: 20px;
279
+ font-size: 0.8rem;
280
+ font-weight: 500;
281
+ margin-right: 0.5rem;
282
+ }
283
+
284
+ .agent-badge.product-manager {
285
+ background: rgba(102, 126, 234, 0.2);
286
+ color: #667eea;
287
+ }
288
+
289
+ .agent-badge.program-manager {
290
+ background: rgba(240, 147, 251, 0.2);
291
+ color: #f093fb;
292
+ }
293
+
294
+ .agent-badge.dev-engineer {
295
+ background: rgba(0, 242, 254, 0.2);
296
+ color: #00f2fe;
297
+ }
298
+
299
+ /* Accordion Styling */
300
+ .accordion {
301
+ border: 1px solid var(--card-border) !important;
302
+ border-radius: 12px !important;
303
+ background: var(--card-bg) !important;
304
+ margin-bottom: 0.5rem !important;
305
+ }
306
+
307
+ /* File Upload */
308
+ .file-upload {
309
+ border: 2px dashed var(--card-border) !important;
310
+ border-radius: 12px !important;
311
+ background: rgba(0, 0, 0, 0.2) !important;
312
+ transition: all 0.3s ease !important;
313
+ }
314
+
315
+ .file-upload:hover {
316
+ border-color: var(--accent-purple) !important;
317
+ background: rgba(102, 126, 234, 0.05) !important;
318
+ }
319
+
320
+ /* Scrollbar */
321
+ ::-webkit-scrollbar {
322
+ width: 8px;
323
+ height: 8px;
324
+ }
325
+
326
+ ::-webkit-scrollbar-track {
327
+ background: var(--dark-bg);
328
+ }
329
+
330
+ ::-webkit-scrollbar-thumb {
331
+ background: var(--card-border);
332
+ border-radius: 4px;
333
+ }
334
+
335
+ ::-webkit-scrollbar-thumb:hover {
336
+ background: var(--accent-purple);
337
+ }
338
+
339
+ /* Loading Animation */
340
+ .loading-dots {
341
+ display: inline-flex;
342
+ gap: 4px;
343
+ }
344
+
345
+ .loading-dots span {
346
+ width: 8px;
347
+ height: 8px;
348
+ border-radius: 50%;
349
+ background: var(--accent-purple);
350
+ animation: bounce 1.4s infinite ease-in-out both;
351
+ }
352
+
353
+ .loading-dots span:nth-child(1) { animation-delay: -0.32s; }
354
+ .loading-dots span:nth-child(2) { animation-delay: -0.16s; }
355
+
356
+ @keyframes bounce {
357
+ 0%, 80%, 100% { transform: scale(0); }
358
+ 40% { transform: scale(1); }
359
+ }
360
+
361
+ /* Progress Bar */
362
+ .progress-bar {
363
+ height: 4px;
364
+ background: var(--card-border);
365
+ border-radius: 2px;
366
+ overflow: hidden;
367
+ margin-top: 0.5rem;
368
+ }
369
+
370
+ .progress-bar-fill {
371
+ height: 100%;
372
+ background: var(--primary-gradient);
373
+ border-radius: 2px;
374
+ transition: width 0.5s ease;
375
+ }
376
+
377
+ /* Example Card */
378
+ .example-card {
379
+ background: rgba(255, 255, 255, 0.02);
380
+ border: 1px solid var(--card-border);
381
+ border-radius: 12px;
382
+ padding: 1rem;
383
+ cursor: pointer;
384
+ transition: all 0.3s ease;
385
+ }
386
+
387
+ .example-card:hover {
388
+ border-color: var(--accent-purple);
389
+ transform: translateY(-2px);
390
+ }
391
+
392
+ .example-title {
393
+ font-weight: 600;
394
+ color: var(--text-primary);
395
+ margin-bottom: 0.5rem;
396
+ }
397
+
398
+ .example-desc {
399
+ font-size: 0.85rem;
400
+ color: var(--text-secondary);
401
+ }
402
+ """
403
+
404
+ # ============================================================================
405
+ # EXAMPLE PRODUCT SPECIFICATIONS
406
+ # ============================================================================
407
+
408
+ EXAMPLE_SPECS = {
409
+ "Email Router": """# Email Router Product Specification
410
+
411
+ ## Executive Summary
412
+ The Email Router is an intelligent email management system designed to automatically classify, route, and respond to incoming emails using advanced NLP and machine learning capabilities.
413
+
414
+ ## Target Users
415
+ - Customer Support Representatives
416
+ - IT Administrators
417
+ - Subject Matter Experts (SMEs)
418
+ - Business Operations Teams
419
+
420
+ ## Core Features
421
+
422
+ ### 1. Intelligent Email Classification
423
+ - Automatic categorization of incoming emails by intent, urgency, and topic
424
+ - Support for custom classification rules and machine learning models
425
+ - Real-time processing with sub-second latency requirements
426
+
427
+ ### 2. Smart Routing Engine
428
+ - Rule-based routing to appropriate departments or individuals
429
+ - Load balancing across support teams
430
+ - Escalation workflows for high-priority items
431
+ - Integration with existing ticketing systems
432
+
433
+ ### 3. Automated Response System
434
+ - Template-based auto-replies for common inquiries
435
+ - AI-generated response suggestions for complex queries
436
+ - Sentiment analysis to adjust tone appropriately
437
+ - Multi-language support
438
+
439
+ ### 4. Analytics Dashboard
440
+ - Real-time metrics on email volume, response times, and resolution rates
441
+ - Historical trend analysis and forecasting
442
+ - Agent performance tracking
443
+ - Customer satisfaction scoring
444
+
445
+ ### 5. Administration Console
446
+ - User and role management
447
+ - Routing rule configuration
448
+ - Knowledge base management
449
+ - System health monitoring
450
+
451
+ ## Technical Requirements
452
+ - Process 10,000+ emails per hour
453
+ - 99.9% uptime SLA
454
+ - Integration via SMTP, IMAP, and REST APIs
455
+ - SOC 2 Type II compliance
456
+ - GDPR and CCPA data handling
457
+
458
+ ## Success Metrics
459
+ - 70% reduction in manual email triage time
460
+ - 50% improvement in first-response time
461
+ - 90% customer satisfaction score
462
+ - 95% routing accuracy
463
+ """,
464
+
465
+ "Task Management App": """# Task Management Application Specification
466
+
467
+ ## Product Vision
468
+ A modern, collaborative task management platform that helps teams organize work, track progress, and meet deadlines efficiently.
469
+
470
+ ## Target Users
471
+ - Project Managers
472
+ - Team Members
473
+ - Executives/Stakeholders
474
+ - Remote Teams
475
+
476
+ ## Core Features
477
+
478
+ ### 1. Task Creation & Organization
479
+ - Create tasks with titles, descriptions, due dates, and priorities
480
+ - Organize tasks into projects and workspaces
481
+ - Support for subtasks and checklists
482
+ - File attachments and rich text descriptions
483
+
484
+ ### 2. Collaboration Tools
485
+ - Assign tasks to team members
486
+ - Comment threads on tasks
487
+ - @mentions and notifications
488
+ - Real-time updates and sync
489
+
490
+ ### 3. Views & Visualization
491
+ - List view for detailed task management
492
+ - Kanban board for workflow visualization
493
+ - Calendar view for deadline tracking
494
+ - Gantt chart for project timeline
495
+
496
+ ### 4. Automation
497
+ - Recurring task creation
498
+ - Due date reminders
499
+ - Status change triggers
500
+ - Integration webhooks
501
+
502
+ ### 5. Reporting & Analytics
503
+ - Project progress dashboards
504
+ - Team workload distribution
505
+ - Time tracking and estimates
506
+ - Custom report generation
507
+
508
+ ## Technical Requirements
509
+ - Web, iOS, and Android applications
510
+ - Real-time synchronization
511
+ - Offline mode support
512
+ - Third-party integrations (Slack, GitHub, Google Calendar)
513
+ - Enterprise SSO support
514
+
515
+ ## Success Metrics
516
+ - 30% improvement in project delivery time
517
+ - 50% reduction in missed deadlines
518
+ - 85% daily active user rate
519
+ - 4.5+ app store rating
520
+ """,
521
+
522
+ "E-Commerce Platform": """# E-Commerce Platform Specification
523
+
524
+ ## Product Overview
525
+ A scalable e-commerce platform enabling businesses to create online stores, manage inventory, process payments, and fulfill orders.
526
+
527
+ ## Target Users
528
+ - Small Business Owners
529
+ - E-commerce Managers
530
+ - Warehouse Staff
531
+ - Customer Service Teams
532
+
533
+ ## Core Features
534
+
535
+ ### 1. Store Management
536
+ - Customizable storefront templates
537
+ - Product catalog management
538
+ - Category and collection organization
539
+ - SEO optimization tools
540
+
541
+ ### 2. Inventory System
542
+ - Real-time stock tracking
543
+ - Multi-warehouse support
544
+ - Low stock alerts
545
+ - Supplier management
546
+
547
+ ### 3. Order Processing
548
+ - Shopping cart functionality
549
+ - Multiple payment gateways
550
+ - Order status tracking
551
+ - Return and refund handling
552
+
553
+ ### 4. Customer Management
554
+ - Customer accounts and profiles
555
+ - Order history
556
+ - Wishlist functionality
557
+ - Loyalty program support
558
+
559
+ ### 5. Analytics & Reports
560
+ - Sales dashboards
561
+ - Customer behavior analytics
562
+ - Inventory reports
563
+ - Financial summaries
564
+
565
+ ## Technical Requirements
566
+ - Support 100,000+ concurrent users
567
+ - PCI DSS compliance
568
+ - Mobile-responsive design
569
+ - API for third-party integrations
570
+ - Multi-currency and multi-language
571
+
572
+ ## Success Metrics
573
+ - 99.99% checkout availability
574
+ - <2 second page load time
575
+ - 3% conversion rate
576
+ - <1% cart abandonment from technical issues
577
+ """
578
+ }
579
+
580
+ # ============================================================================
581
+ # WORKFLOW KNOWLEDGE DEFINITIONS
582
+ # ============================================================================
583
+
584
+ KNOWLEDGE_ACTION_PLANNING = """
585
+ Stories are defined from a product spec by identifying a persona, an action, and a desired outcome for each story.
586
+ Each story represents a specific functionality of the product described in the specification.
587
+ Features are defined by grouping related user stories into cohesive capabilities.
588
+ Tasks are defined for each story and represent the engineering work required to develop the product.
589
+ A development Plan for a product contains all these components organized in a logical structure.
590
+ """
591
+
592
+ PERSONA_PRODUCT_MANAGER = "You are a Product Manager responsible for defining user stories for a product."
593
+ PERSONA_PROGRAM_MANAGER = "You are a Program Manager responsible for defining product features."
594
+ PERSONA_DEV_ENGINEER = "You are a Development Engineer responsible for defining engineering tasks."
595
+
596
+ EVALUATION_CRITERIA_USER_STORIES = (
597
+ "The answer should be stories that follow the following structure: "
598
+ "As a [type of user], I want [an action or feature] so that [benefit/value]."
599
+ )
600
+
601
+ EVALUATION_CRITERIA_FEATURES = (
602
+ "The answer should be product features that follow the following structure: "
603
+ "Feature Name: A clear, concise title that identifies the capability\n"
604
+ "Description: A brief explanation of what the feature does and its purpose\n"
605
+ "Key Functionality: The specific capabilities or actions the feature provides\n"
606
+ "User Benefit: How this feature creates value for the user"
607
+ )
608
+
609
+ EVALUATION_CRITERIA_TASKS = (
610
+ "The answer should be tasks following this exact structure: "
611
+ "Task ID: A unique identifier for tracking purposes\n"
612
+ "Task Title: Brief description of the specific development work\n"
613
+ "Related User Story: Reference to the parent user story\n"
614
+ "Description: Detailed explanation of the technical work required\n"
615
+ "Acceptance Criteria: Specific requirements that must be met for completion\n"
616
+ "Estimated Effort: Time or complexity estimation\n"
617
+ "Dependencies: Any tasks that must be completed first"
618
+ )
619
+
620
+ # ============================================================================
621
+ # WORKFLOW ENGINE
622
+ # ============================================================================
623
+
624
+ class ProjectPlannerWorkflow:
625
+ """
626
+ Orchestrates the multi-agent workflow for generating project plans.
627
+ """
628
+
629
+ def __init__(self, api_key: str):
630
+ self.api_key = api_key
631
+ self.agents_initialized = False
632
+ self.workflow_log = []
633
+ self.stats = {
634
+ "user_stories": 0,
635
+ "features": 0,
636
+ "tasks": 0,
637
+ "iterations": 0
638
+ }
639
+
640
+ def log(self, message: str, level: str = "info"):
641
+ """Add a message to the workflow log."""
642
+ timestamp = datetime.now().strftime("%H:%M:%S")
643
+ self.workflow_log.append({
644
+ "timestamp": timestamp,
645
+ "level": level,
646
+ "message": message
647
+ })
648
+
649
+ def initialize_agents(self, product_spec: str):
650
+ """Initialize all agents with the product specification."""
651
+ self.log("Initializing AI agents...")
652
+
653
+ # Action Planning Agent
654
+ self.action_planning_agent = ActionPlanningAgent(
655
+ self.api_key,
656
+ KNOWLEDGE_ACTION_PLANNING
657
+ )
658
+
659
+ # Product Manager Team
660
+ knowledge_pm = (
661
+ "Stories are defined by writing sentences with a persona, an action, and a desired outcome. "
662
+ "The sentences always start with: As a "
663
+ "Write several stories for the product spec below, where the personas are the different users of the product. "
664
+ f"\n\nProduct Specification:\n{product_spec}"
665
+ )
666
+
667
+ self.product_manager_agent = KnowledgeAugmentedPromptAgent(
668
+ self.api_key,
669
+ PERSONA_PRODUCT_MANAGER,
670
+ knowledge_pm
671
+ )
672
+
673
+ self.product_manager_evaluator = EvaluationAgent(
674
+ openai_api_key=self.api_key,
675
+ persona="You are an evaluation agent that checks the answers of other worker agents",
676
+ evaluation_criteria=EVALUATION_CRITERIA_USER_STORIES,
677
+ worker_agent=self.product_manager_agent,
678
+ max_interactions=5
679
+ )
680
+
681
+ # Program Manager Team
682
+ knowledge_pgm = (
683
+ "Features of a product are defined by organizing similar user stories into cohesive groups. "
684
+ f"\n\nProduct Specification:\n{product_spec}"
685
+ )
686
+
687
+ self.program_manager_agent = KnowledgeAugmentedPromptAgent(
688
+ self.api_key,
689
+ PERSONA_PROGRAM_MANAGER,
690
+ knowledge_pgm
691
+ )
692
+
693
+ self.program_manager_evaluator = EvaluationAgent(
694
+ openai_api_key=self.api_key,
695
+ persona="You are an evaluation agent that checks the answers of other worker agents",
696
+ evaluation_criteria=EVALUATION_CRITERIA_FEATURES,
697
+ worker_agent=self.program_manager_agent,
698
+ max_interactions=5
699
+ )
700
+
701
+ # Development Engineer Team
702
+ knowledge_dev = (
703
+ "Development tasks are defined by identifying what needs to be built to implement each user story. "
704
+ f"\n\nProduct Specification:\n{product_spec}"
705
+ )
706
+
707
+ self.dev_engineer_agent = KnowledgeAugmentedPromptAgent(
708
+ self.api_key,
709
+ PERSONA_DEV_ENGINEER,
710
+ knowledge_dev
711
+ )
712
+
713
+ self.dev_engineer_evaluator = EvaluationAgent(
714
+ openai_api_key=self.api_key,
715
+ persona="You are an evaluation agent that checks the answers of other worker agents",
716
+ evaluation_criteria=EVALUATION_CRITERIA_TASKS,
717
+ worker_agent=self.dev_engineer_agent,
718
+ max_interactions=5
719
+ )
720
+
721
+ # Routing Agent
722
+ routes = [
723
+ {
724
+ "name": "Product Manager",
725
+ "description": "Responsible for defining product personas and user stories only.",
726
+ "func": self._product_manager_handler
727
+ },
728
+ {
729
+ "name": "Program Manager",
730
+ "description": "Responsible for defining product features by grouping related user stories.",
731
+ "func": self._program_manager_handler
732
+ },
733
+ {
734
+ "name": "Development Engineer",
735
+ "description": "Responsible for defining development tasks and engineering work.",
736
+ "func": self._dev_engineer_handler
737
+ }
738
+ ]
739
+
740
+ self.routing_agent = RoutingAgent(self.api_key, routes)
741
+
742
+ self.agents_initialized = True
743
+ self.log("All agents initialized successfully", "success")
744
+
745
+ def _product_manager_handler(self, query: str) -> str:
746
+ """Handle Product Manager routing."""
747
+ self.log("Product Manager generating user stories...")
748
+ response = self.product_manager_agent.respond(query)
749
+ result = self.product_manager_evaluator.evaluate(response)
750
+ self.stats["iterations"] += result.get("iterations", 1)
751
+ return result["final_response"]
752
+
753
+ def _program_manager_handler(self, query: str) -> str:
754
+ """Handle Program Manager routing."""
755
+ self.log("Program Manager defining features...")
756
+ response = self.program_manager_agent.respond(query)
757
+ result = self.program_manager_evaluator.evaluate(response)
758
+ self.stats["iterations"] += result.get("iterations", 1)
759
+ return result["final_response"]
760
+
761
+ def _dev_engineer_handler(self, query: str) -> str:
762
+ """Handle Development Engineer routing."""
763
+ self.log("Development Engineer creating tasks...")
764
+ response = self.dev_engineer_agent.respond(query)
765
+ result = self.dev_engineer_evaluator.evaluate(response)
766
+ self.stats["iterations"] += result.get("iterations", 1)
767
+ return result["final_response"]
768
+
769
+ def run_workflow(self, product_spec: str) -> Generator[Tuple[str, str, str, str, str], None, None]:
770
+ """
771
+ Execute the full workflow and yield progress updates.
772
+
773
+ Yields tuples of:
774
+ (workflow_status, user_stories, features, tasks, full_plan)
775
+ """
776
+ self.workflow_log = []
777
+ self.stats = {"user_stories": 0, "features": 0, "tasks": 0, "iterations": 0}
778
+
779
+ # Initialize agents
780
+ yield ("⏳ Initializing agents...", "", "", "", "")
781
+ self.initialize_agents(product_spec)
782
+
783
+ # Extract workflow steps
784
+ yield ("🔍 Planning workflow steps...", "", "", "", "")
785
+ self.log("Action Planning Agent analyzing product specification...")
786
+
787
+ workflow_prompt = (
788
+ "Create a comprehensive project plan including: "
789
+ "1) User stories for all personas, "
790
+ "2) Key product features grouped logically, "
791
+ "3) Development tasks for implementation"
792
+ )
793
+
794
+ steps = self.action_planning_agent.extract_steps_from_prompt(workflow_prompt)
795
+ self.log(f"Identified {len(steps)} workflow steps", "success")
796
+
797
+ # Process each step
798
+ user_stories = ""
799
+ features = ""
800
+ tasks = ""
801
+
802
+ # Generate User Stories
803
+ yield ("👤 Generating user stories...", "", "", "", "")
804
+ self.log("Routing to Product Manager team...")
805
+
806
+ user_stories_prompt = "Define comprehensive user stories for all personas in the product specification"
807
+ user_stories = self._product_manager_handler(user_stories_prompt)
808
+ self.stats["user_stories"] = user_stories.count("As a")
809
+ self.log(f"Generated {self.stats['user_stories']} user stories", "success")
810
+
811
+ yield ("✅ User stories complete", user_stories, "", "", "")
812
+
813
+ # Generate Features
814
+ yield ("📦 Defining product features...", user_stories, "", "", "")
815
+ self.log("Routing to Program Manager team...")
816
+
817
+ features_prompt = f"Based on these user stories, define the key product features:\n\n{user_stories}"
818
+ features = self._program_manager_handler(features_prompt)
819
+ self.stats["features"] = features.lower().count("feature name")
820
+ self.log(f"Defined {self.stats['features']} product features", "success")
821
+
822
+ yield ("✅ Features complete", user_stories, features, "", "")
823
+
824
+ # Generate Tasks
825
+ yield ("🔧 Creating engineering tasks...", user_stories, features, "", "")
826
+ self.log("Routing to Development Engineer team...")
827
+
828
+ tasks_prompt = f"Create detailed engineering tasks to implement these features:\n\n{features}"
829
+ tasks = self._dev_engineer_handler(tasks_prompt)
830
+ self.stats["tasks"] = tasks.lower().count("task id")
831
+ self.log(f"Created {self.stats['tasks']} engineering tasks", "success")
832
+
833
+ yield ("✅ Tasks complete", user_stories, features, tasks, "")
834
+
835
+ # Generate Full Plan
836
+ yield ("📋 Compiling final plan...", user_stories, features, tasks, "")
837
+
838
+ full_plan = self._compile_full_plan(user_stories, features, tasks)
839
+ self.log("Project plan compilation complete!", "success")
840
+
841
+ yield ("✅ Project plan ready!", user_stories, features, tasks, full_plan)
842
+
843
+ def _compile_full_plan(self, user_stories: str, features: str, tasks: str) -> str:
844
+ """Compile all outputs into a formatted project plan."""
845
+ plan = f"""# 📋 Project Plan
846
+
847
+ ## Generated: {datetime.now().strftime("%Y-%m-%d %H:%M")}
848
+
849
+ ---
850
+
851
+ ## 📊 Summary Statistics
852
+
853
+ | Metric | Count |
854
+ |--------|-------|
855
+ | User Stories | {self.stats['user_stories']} |
856
+ | Product Features | {self.stats['features']} |
857
+ | Engineering Tasks | {self.stats['tasks']} |
858
+ | Evaluation Iterations | {self.stats['iterations']} |
859
+
860
+ ---
861
+
862
+ ## 👤 User Stories
863
+
864
+ {user_stories}
865
+
866
+ ---
867
+
868
+ ## 📦 Product Features
869
+
870
+ {features}
871
+
872
+ ---
873
+
874
+ ## 🔧 Engineering Tasks
875
+
876
+ {tasks}
877
+
878
+ ---
879
+
880
+ ## 📝 Workflow Log
881
+
882
+ | Time | Level | Message |
883
+ |------|-------|---------|
884
+ """
885
+ for log_entry in self.workflow_log:
886
+ plan += f"| {log_entry['timestamp']} | {log_entry['level']} | {log_entry['message']} |\n"
887
+
888
+ return plan
889
+
890
+
891
+ # ============================================================================
892
+ # GRADIO INTERFACE
893
+ # ============================================================================
894
+
895
+ def create_interface():
896
+ """Create the Gradio interface."""
897
+
898
+ workflow = None
899
+
900
+ def validate_api_key(api_key: str) -> str:
901
+ """Validate the API key format."""
902
+ if not api_key or len(api_key) < 20:
903
+ return "❌ Please enter a valid API key"
904
+ return "✅ API key accepted"
905
+
906
+ def load_example(example_name: str) -> str:
907
+ """Load an example product specification."""
908
+ return EXAMPLE_SPECS.get(example_name, "")
909
+
910
+ def run_planner(
911
+ api_key: str,
912
+ product_spec: str,
913
+ progress=gr.Progress()
914
+ ) -> Tuple[str, str, str, str, str]:
915
+ """Run the project planner workflow."""
916
+
917
+ if not api_key or len(api_key) < 20:
918
+ return (
919
+ "❌ Error: Please enter a valid OpenAI API key",
920
+ "", "", "", ""
921
+ )
922
+
923
+ if not product_spec or len(product_spec) < 100:
924
+ return (
925
+ "❌ Error: Please provide a product specification (minimum 100 characters)",
926
+ "", "", "", ""
927
+ )
928
+
929
+ try:
930
+ planner = ProjectPlannerWorkflow(api_key)
931
+
932
+ results = ("", "", "", "", "")
933
+ progress(0, desc="Starting workflow...")
934
+
935
+ step_progress = [0.1, 0.2, 0.4, 0.6, 0.8, 0.9, 1.0]
936
+ step_idx = 0
937
+
938
+ for status, stories, features, tasks, plan in planner.run_workflow(product_spec):
939
+ if step_idx < len(step_progress):
940
+ progress(step_progress[step_idx], desc=status)
941
+ step_idx += 1
942
+ results = (status, stories, features, tasks, plan)
943
+
944
+ return results
945
+
946
+ except Exception as e:
947
+ return (
948
+ f"❌ Error: {str(e)}",
949
+ "", "", "", ""
950
+ )
951
+
952
+ def export_markdown(user_stories: str, features: str, tasks: str) -> str:
953
+ """Export results as markdown."""
954
+ if not any([user_stories, features, tasks]):
955
+ return None
956
+
957
+ content = f"""# Project Plan Export
958
+ Generated: {datetime.now().strftime("%Y-%m-%d %H:%M")}
959
+
960
+ ## User Stories
961
+ {user_stories}
962
+
963
+ ## Product Features
964
+ {features}
965
+
966
+ ## Engineering Tasks
967
+ {tasks}
968
+ """
969
+ # Save to temp file
970
+ filename = f"project_plan_{datetime.now().strftime('%Y%m%d_%H%M%S')}.md"
971
+ with open(filename, "w") as f:
972
+ f.write(content)
973
+ return filename
974
+
975
+ def export_json(user_stories: str, features: str, tasks: str) -> str:
976
+ """Export results as JSON."""
977
+ if not any([user_stories, features, tasks]):
978
+ return None
979
+
980
+ data = {
981
+ "generated_at": datetime.now().isoformat(),
982
+ "user_stories": user_stories,
983
+ "features": features,
984
+ "tasks": tasks
985
+ }
986
+
987
+ filename = f"project_plan_{datetime.now().strftime('%Y%m%d_%H%M%S')}.json"
988
+ with open(filename, "w") as f:
989
+ json.dump(data, f, indent=2)
990
+ return filename
991
+
992
+ # Build the interface
993
+ with gr.Blocks(css=CUSTOM_CSS, title="AI Project Planner", theme=gr.themes.Base()) as app:
994
+
995
+ # Header
996
+ gr.HTML("""
997
+ <div class="header-container">
998
+ <div class="header-title">🚀 AI Project Planner</div>
999
+ <div class="header-subtitle">
1000
+ Transform product specifications into comprehensive project plans using orchestrated AI agents
1001
+ </div>
1002
+ </div>
1003
+ """)
1004
+
1005
+ with gr.Row():
1006
+ # Left Column - Input
1007
+ with gr.Column(scale=1):
1008
+ gr.Markdown("### ⚙️ Configuration")
1009
+
1010
+ api_key_input = gr.Textbox(
1011
+ label="OpenAI API Key",
1012
+ placeholder="sk-...",
1013
+ type="password",
1014
+ info="Your API key is not stored and only used for this session"
1015
+ )
1016
+
1017
+ api_status = gr.Markdown("Enter your API key to begin")
1018
+
1019
+ api_key_input.change(
1020
+ fn=validate_api_key,
1021
+ inputs=[api_key_input],
1022
+ outputs=[api_status]
1023
+ )
1024
+
1025
+ gr.Markdown("### 📄 Product Specification")
1026
+
1027
+ with gr.Accordion("📚 Load Example", open=False):
1028
+ example_dropdown = gr.Dropdown(
1029
+ choices=list(EXAMPLE_SPECS.keys()),
1030
+ label="Select Example",
1031
+ value="Email Router"
1032
+ )
1033
+ load_example_btn = gr.Button("Load Example", variant="secondary")
1034
+
1035
+ spec_input = gr.Textbox(
1036
+ label="Product Specification",
1037
+ placeholder="Paste your product specification here...",
1038
+ lines=15,
1039
+ max_lines=30
1040
+ )
1041
+
1042
+ load_example_btn.click(
1043
+ fn=load_example,
1044
+ inputs=[example_dropdown],
1045
+ outputs=[spec_input]
1046
+ )
1047
+
1048
+ generate_btn = gr.Button(
1049
+ "🎯 Generate Project Plan",
1050
+ variant="primary",
1051
+ elem_classes=["primary-btn"]
1052
+ )
1053
+
1054
+ # Right Column - Output
1055
+ with gr.Column(scale=2):
1056
+ gr.Markdown("### 📊 Results")
1057
+
1058
+ workflow_status = gr.Markdown("Ready to generate project plan")
1059
+
1060
+ with gr.Tabs():
1061
+ with gr.Tab("👤 User Stories"):
1062
+ user_stories_output = gr.Markdown(
1063
+ label="User Stories",
1064
+ elem_classes=["output-content"]
1065
+ )
1066
+
1067
+ with gr.Tab("📦 Features"):
1068
+ features_output = gr.Markdown(
1069
+ label="Product Features",
1070
+ elem_classes=["output-content"]
1071
+ )
1072
+
1073
+ with gr.Tab("🔧 Tasks"):
1074
+ tasks_output = gr.Markdown(
1075
+ label="Engineering Tasks",
1076
+ elem_classes=["output-content"]
1077
+ )
1078
+
1079
+ with gr.Tab("📋 Full Plan"):
1080
+ full_plan_output = gr.Markdown(
1081
+ label="Complete Project Plan",
1082
+ elem_classes=["output-content"]
1083
+ )
1084
+
1085
+ gr.Markdown("### 📥 Export")
1086
+
1087
+ with gr.Row():
1088
+ export_md_btn = gr.Button("📄 Download Markdown", variant="secondary")
1089
+ export_json_btn = gr.Button("📊 Download JSON", variant="secondary")
1090
+
1091
+ export_file = gr.File(label="Download", visible=False)
1092
+
1093
+ # Footer
1094
+ gr.HTML("""
1095
+ <div style="text-align: center; margin-top: 2rem; padding: 1rem; border-top: 1px solid #2d2d44;">
1096
+ <p style="color: #a0a0b0; font-size: 0.9rem;">
1097
+ Built with ❤️ using Gradio • Powered by OpenAI GPT-3.5 •
1098
+ <a href="https://github.com" style="color: #667eea;">View on GitHub</a>
1099
+ </p>
1100
+ </div>
1101
+ """)
1102
+
1103
+ # Event handlers
1104
+ generate_btn.click(
1105
+ fn=run_planner,
1106
+ inputs=[api_key_input, spec_input],
1107
+ outputs=[
1108
+ workflow_status,
1109
+ user_stories_output,
1110
+ features_output,
1111
+ tasks_output,
1112
+ full_plan_output
1113
+ ],
1114
+ show_progress=True
1115
+ )
1116
+
1117
+ export_md_btn.click(
1118
+ fn=export_markdown,
1119
+ inputs=[user_stories_output, features_output, tasks_output],
1120
+ outputs=[export_file]
1121
+ )
1122
+
1123
+ export_json_btn.click(
1124
+ fn=export_json,
1125
+ inputs=[user_stories_output, features_output, tasks_output],
1126
+ outputs=[export_file]
1127
+ )
1128
+
1129
+ return app
1130
+
1131
+
1132
+ # ============================================================================
1133
+ # MAIN ENTRY POINT
1134
+ # ============================================================================
1135
+
1136
+ if __name__ == "__main__":
1137
+ app = create_interface()
1138
+ app.launch(
1139
+ server_name="0.0.0.0",
1140
+ server_port=7860,
1141
+ share=False,
1142
+ show_error=True
1143
+ )
base_agents.py ADDED
@@ -0,0 +1,525 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ """
2
+ Base Agents Library for AI Project Planner
3
+
4
+ A collection of reusable AI agents for building agentic workflows.
5
+ Optimized for web deployment with proper error handling and logging.
6
+ """
7
+
8
+ from openai import OpenAI
9
+ import numpy as np
10
+ from typing import Dict, Any, List, Optional, Callable
11
+ import logging
12
+
13
+ # Configure logging
14
+ logging.basicConfig(level=logging.INFO)
15
+ logger = logging.getLogger(__name__)
16
+
17
+
18
+ class DirectPromptAgent:
19
+ """
20
+ A Direct Prompt Agent that relays user input directly to the LLM
21
+ without additional context, memory, or specialized tools.
22
+
23
+ This is the most straightforward method for interacting with an LLM.
24
+ """
25
+
26
+ def __init__(self, openai_api_key: str, base_url: str = None):
27
+ """
28
+ Initialize the DirectPromptAgent.
29
+
30
+ Args:
31
+ openai_api_key: The OpenAI API key for authentication.
32
+ base_url: Optional custom base URL for API calls.
33
+ """
34
+ self.openai_api_key = openai_api_key
35
+ self.base_url = base_url
36
+
37
+ def _get_client(self) -> OpenAI:
38
+ """Create and return an OpenAI client instance."""
39
+ if self.base_url:
40
+ return OpenAI(api_key=self.openai_api_key, base_url=self.base_url)
41
+ return OpenAI(api_key=self.openai_api_key)
42
+
43
+ def respond(self, prompt: str) -> str:
44
+ """
45
+ Generate a response by passing the user prompt directly to the LLM.
46
+
47
+ Args:
48
+ prompt: The user's input prompt.
49
+
50
+ Returns:
51
+ The textual content of the LLM's response.
52
+ """
53
+ try:
54
+ client = self._get_client()
55
+
56
+ response = client.chat.completions.create(
57
+ model="gpt-3.5-turbo",
58
+ messages=[
59
+ {"role": "user", "content": prompt}
60
+ ],
61
+ temperature=0
62
+ )
63
+
64
+ return response.choices[0].message.content
65
+
66
+ except Exception as e:
67
+ logger.error(f"DirectPromptAgent error: {str(e)}")
68
+ raise
69
+
70
+
71
+ class AugmentedPromptAgent:
72
+ """
73
+ An Augmented Prompt Agent that responds according to a predefined persona.
74
+
75
+ This agent explicitly adopts a persona via a system prompt, leading to
76
+ more targeted and contextually relevant outputs.
77
+ """
78
+
79
+ def __init__(self, openai_api_key: str, persona: str, base_url: str = None):
80
+ """
81
+ Initialize the AugmentedPromptAgent.
82
+
83
+ Args:
84
+ openai_api_key: The OpenAI API key for authentication.
85
+ persona: The persona the agent should assume.
86
+ base_url: Optional custom base URL for API calls.
87
+ """
88
+ self.openai_api_key = openai_api_key
89
+ self.persona = persona
90
+ self.base_url = base_url
91
+
92
+ def _get_client(self) -> OpenAI:
93
+ """Create and return an OpenAI client instance."""
94
+ if self.base_url:
95
+ return OpenAI(api_key=self.openai_api_key, base_url=self.base_url)
96
+ return OpenAI(api_key=self.openai_api_key)
97
+
98
+ def respond(self, input_text: str) -> str:
99
+ """
100
+ Generate a response using the defined persona.
101
+
102
+ Args:
103
+ input_text: The user's input prompt.
104
+
105
+ Returns:
106
+ The textual content of the LLM's response.
107
+ """
108
+ try:
109
+ client = self._get_client()
110
+
111
+ system_prompt = f"{self.persona} Forget all previous conversational context."
112
+
113
+ response = client.chat.completions.create(
114
+ model="gpt-3.5-turbo",
115
+ messages=[
116
+ {"role": "system", "content": system_prompt},
117
+ {"role": "user", "content": input_text}
118
+ ],
119
+ temperature=1.0
120
+ )
121
+
122
+ return response.choices[0].message.content
123
+
124
+ except Exception as e:
125
+ logger.error(f"AugmentedPromptAgent error: {str(e)}")
126
+ raise
127
+
128
+
129
+ class KnowledgeAugmentedPromptAgent:
130
+ """
131
+ A Knowledge Augmented Prompt Agent that incorporates specific knowledge
132
+ alongside a defined persona when responding to prompts.
133
+
134
+ This agent ensures answers are based on explicitly provided information
135
+ rather than the LLM's inherent knowledge.
136
+ """
137
+
138
+ def __init__(
139
+ self,
140
+ openai_api_key: str,
141
+ persona: str,
142
+ knowledge: str,
143
+ base_url: str = None
144
+ ):
145
+ """
146
+ Initialize the KnowledgeAugmentedPromptAgent.
147
+
148
+ Args:
149
+ openai_api_key: The OpenAI API key for authentication.
150
+ persona: The persona the agent should assume.
151
+ knowledge: The specific knowledge to use for responses.
152
+ base_url: Optional custom base URL for API calls.
153
+ """
154
+ self.openai_api_key = openai_api_key
155
+ self.persona = persona
156
+ self.knowledge = knowledge
157
+ self.base_url = base_url
158
+
159
+ def _get_client(self) -> OpenAI:
160
+ """Create and return an OpenAI client instance."""
161
+ if self.base_url:
162
+ return OpenAI(api_key=self.openai_api_key, base_url=self.base_url)
163
+ return OpenAI(api_key=self.openai_api_key)
164
+
165
+ def respond(self, input_text: str) -> str:
166
+ """
167
+ Generate a response using the persona and provided knowledge only.
168
+
169
+ Args:
170
+ input_text: The user's input prompt.
171
+
172
+ Returns:
173
+ The textual content of the LLM's response.
174
+ """
175
+ try:
176
+ client = self._get_client()
177
+
178
+ system_prompt = (
179
+ f"You are {self.persona} knowledge-based assistant. Forget all previous context. "
180
+ f"Use only the following knowledge to answer, do not use your own knowledge: {self.knowledge} "
181
+ f"Answer the prompt based on this knowledge, not your own."
182
+ )
183
+
184
+ response = client.chat.completions.create(
185
+ model="gpt-3.5-turbo",
186
+ messages=[
187
+ {"role": "system", "content": system_prompt},
188
+ {"role": "user", "content": input_text}
189
+ ],
190
+ temperature=0
191
+ )
192
+
193
+ return response.choices[0].message.content
194
+
195
+ except Exception as e:
196
+ logger.error(f"KnowledgeAugmentedPromptAgent error: {str(e)}")
197
+ raise
198
+
199
+
200
+ class EvaluationAgent:
201
+ """
202
+ An Evaluation Agent that assesses responses from a worker agent
203
+ against given criteria, potentially refining through iterative feedback.
204
+
205
+ This agent implements a feedback loop to ensure quality outputs.
206
+ """
207
+
208
+ def __init__(
209
+ self,
210
+ openai_api_key: str,
211
+ persona: str,
212
+ evaluation_criteria: str,
213
+ worker_agent: Any,
214
+ max_interactions: int = 10,
215
+ base_url: str = None
216
+ ):
217
+ """
218
+ Initialize the EvaluationAgent.
219
+
220
+ Args:
221
+ openai_api_key: The OpenAI API key for authentication.
222
+ persona: The evaluator's persona.
223
+ evaluation_criteria: Criteria for evaluating responses.
224
+ worker_agent: The agent whose responses will be evaluated.
225
+ max_interactions: Maximum refinement iterations.
226
+ base_url: Optional custom base URL for API calls.
227
+ """
228
+ self.openai_api_key = openai_api_key
229
+ self.persona = persona
230
+ self.evaluation_criteria = evaluation_criteria
231
+ self.worker_agent = worker_agent
232
+ self.max_interactions = max_interactions
233
+ self.base_url = base_url
234
+
235
+ def _get_client(self) -> OpenAI:
236
+ """Create and return an OpenAI client instance."""
237
+ if self.base_url:
238
+ return OpenAI(api_key=self.openai_api_key, base_url=self.base_url)
239
+ return OpenAI(api_key=self.openai_api_key)
240
+
241
+ def evaluate(self, prompt_to_evaluate: str) -> Dict[str, Any]:
242
+ """
243
+ Evaluate and potentially refine a response through iteration.
244
+
245
+ Args:
246
+ prompt_to_evaluate: The initial prompt to evaluate.
247
+
248
+ Returns:
249
+ Dictionary containing final_response, evaluation, and iterations.
250
+ """
251
+ try:
252
+ client = self._get_client()
253
+ initial_prompt = prompt_to_evaluate
254
+ response_from_worker = ""
255
+ evaluation = ""
256
+
257
+ for i in range(self.max_interactions):
258
+ logger.info(f"Evaluation iteration {i + 1}/{self.max_interactions}")
259
+
260
+ # Step 1: Get response from worker agent
261
+ response_from_worker = self.worker_agent.respond(prompt_to_evaluate)
262
+
263
+ # Step 2: Evaluate the response
264
+ eval_prompt = (
265
+ f"Does the following answer: {response_from_worker}\n"
266
+ f"Meet this criteria: {self.evaluation_criteria}\n"
267
+ f"Respond Yes or No, and the reason why it does or doesn't meet the criteria."
268
+ )
269
+
270
+ response = client.chat.completions.create(
271
+ model="gpt-3.5-turbo",
272
+ messages=[
273
+ {"role": "system", "content": self.persona},
274
+ {"role": "user", "content": eval_prompt}
275
+ ],
276
+ temperature=0
277
+ )
278
+
279
+ evaluation = response.choices[0].message.content.strip()
280
+
281
+ # Step 3: Check if evaluation passes
282
+ if evaluation.lower().startswith("yes"):
283
+ logger.info("Evaluation passed!")
284
+ break
285
+
286
+ # Step 4: Generate correction instructions
287
+ instruction_prompt = (
288
+ f"Provide instructions to fix an answer based on these reasons "
289
+ f"why it is incorrect: {evaluation}"
290
+ )
291
+
292
+ response = client.chat.completions.create(
293
+ model="gpt-3.5-turbo",
294
+ messages=[
295
+ {"role": "system", "content": "You are a helpful assistant that provides clear correction instructions."},
296
+ {"role": "user", "content": instruction_prompt}
297
+ ],
298
+ temperature=0
299
+ )
300
+
301
+ instructions = response.choices[0].message.content.strip()
302
+
303
+ # Step 5: Prepare refined prompt
304
+ prompt_to_evaluate = (
305
+ f"The original prompt was: {initial_prompt}\n"
306
+ f"The response to that prompt was: {response_from_worker}\n"
307
+ f"It has been evaluated as incorrect.\n"
308
+ f"Make only these corrections, do not alter content validity: {instructions}"
309
+ )
310
+
311
+ return {
312
+ "final_response": response_from_worker,
313
+ "evaluation": evaluation,
314
+ "iterations": i + 1
315
+ }
316
+
317
+ except Exception as e:
318
+ logger.error(f"EvaluationAgent error: {str(e)}")
319
+ raise
320
+
321
+
322
+ class RoutingAgent:
323
+ """
324
+ A Routing Agent that directs user prompts to the most appropriate
325
+ specialized agent based on semantic similarity.
326
+
327
+ Uses embeddings to match prompts with agent descriptions.
328
+ """
329
+
330
+ def __init__(
331
+ self,
332
+ openai_api_key: str,
333
+ agents: List[Dict[str, Any]],
334
+ base_url: str = None
335
+ ):
336
+ """
337
+ Initialize the RoutingAgent.
338
+
339
+ Args:
340
+ openai_api_key: The OpenAI API key for authentication.
341
+ agents: List of agent dictionaries with name, description, and func.
342
+ base_url: Optional custom base URL for API calls.
343
+ """
344
+ self.openai_api_key = openai_api_key
345
+ self.agents = agents
346
+ self.base_url = base_url
347
+ self._embedding_cache = {}
348
+
349
+ def _get_client(self) -> OpenAI:
350
+ """Create and return an OpenAI client instance."""
351
+ if self.base_url:
352
+ return OpenAI(api_key=self.openai_api_key, base_url=self.base_url)
353
+ return OpenAI(api_key=self.openai_api_key)
354
+
355
+ def get_embedding(self, text: str) -> List[float]:
356
+ """
357
+ Calculate the embedding of text using the text-embedding-3-large model.
358
+
359
+ Args:
360
+ text: The text to embed.
361
+
362
+ Returns:
363
+ The embedding vector.
364
+ """
365
+ # Check cache first
366
+ if text in self._embedding_cache:
367
+ return self._embedding_cache[text]
368
+
369
+ try:
370
+ client = self._get_client()
371
+
372
+ response = client.embeddings.create(
373
+ model="text-embedding-3-large",
374
+ input=text,
375
+ encoding_format="float"
376
+ )
377
+
378
+ embedding = response.data[0].embedding
379
+ self._embedding_cache[text] = embedding
380
+
381
+ return embedding
382
+
383
+ except Exception as e:
384
+ logger.error(f"Embedding error: {str(e)}")
385
+ raise
386
+
387
+ def route(self, user_input: str) -> Any:
388
+ """
389
+ Route user prompts to the most appropriate agent based on similarity.
390
+
391
+ Args:
392
+ user_input: The user's input prompt.
393
+
394
+ Returns:
395
+ The response from the selected agent's function.
396
+ """
397
+ try:
398
+ input_emb = self.get_embedding(user_input)
399
+ best_agent = None
400
+ best_score = -1
401
+
402
+ for agent in self.agents:
403
+ agent_emb = self.get_embedding(agent["description"])
404
+
405
+ if agent_emb is None:
406
+ continue
407
+
408
+ # Calculate cosine similarity
409
+ similarity = np.dot(input_emb, agent_emb) / (
410
+ np.linalg.norm(input_emb) * np.linalg.norm(agent_emb)
411
+ )
412
+
413
+ logger.info(f"Agent '{agent['name']}' similarity: {similarity:.3f}")
414
+
415
+ if similarity > best_score:
416
+ best_score = similarity
417
+ best_agent = agent
418
+
419
+ if best_agent is None:
420
+ return "Sorry, no suitable agent could be selected."
421
+
422
+ logger.info(f"Selected agent: {best_agent['name']} (score={best_score:.3f})")
423
+
424
+ return best_agent["func"](user_input)
425
+
426
+ except Exception as e:
427
+ logger.error(f"RoutingAgent error: {str(e)}")
428
+ raise
429
+
430
+
431
+ class ActionPlanningAgent:
432
+ """
433
+ An Action Planning Agent that extracts and lists actionable steps
434
+ from a user's prompt based on provided knowledge.
435
+
436
+ Crucial for constructing agentic workflows by breaking down complex tasks.
437
+ """
438
+
439
+ def __init__(
440
+ self,
441
+ openai_api_key: str,
442
+ knowledge: str,
443
+ base_url: str = None
444
+ ):
445
+ """
446
+ Initialize the ActionPlanningAgent.
447
+
448
+ Args:
449
+ openai_api_key: The OpenAI API key for authentication.
450
+ knowledge: The knowledge base for extracting action steps.
451
+ base_url: Optional custom base URL for API calls.
452
+ """
453
+ self.openai_api_key = openai_api_key
454
+ self.knowledge = knowledge
455
+ self.base_url = base_url
456
+
457
+ def _get_client(self) -> OpenAI:
458
+ """Create and return an OpenAI client instance."""
459
+ if self.base_url:
460
+ return OpenAI(api_key=self.openai_api_key, base_url=self.base_url)
461
+ return OpenAI(api_key=self.openai_api_key)
462
+
463
+ def extract_steps_from_prompt(self, prompt: str) -> List[str]:
464
+ """
465
+ Extract actionable steps from a user prompt using provided knowledge.
466
+
467
+ Args:
468
+ prompt: The user's input prompt describing a task.
469
+
470
+ Returns:
471
+ A clean list of action steps.
472
+ """
473
+ try:
474
+ client = self._get_client()
475
+
476
+ system_prompt = (
477
+ f"You are an action planning agent. Using your knowledge, you extract from the user prompt "
478
+ f"the steps requested to complete the action the user is asking for. You return the steps as a list. "
479
+ f"Only return the steps in your knowledge. Forget any previous context. "
480
+ f"This is your knowledge: {self.knowledge}"
481
+ )
482
+
483
+ response = client.chat.completions.create(
484
+ model="gpt-3.5-turbo",
485
+ messages=[
486
+ {"role": "system", "content": system_prompt},
487
+ {"role": "user", "content": prompt}
488
+ ],
489
+ temperature=0
490
+ )
491
+
492
+ response_text = response.choices[0].message.content
493
+
494
+ # Clean and format the extracted steps
495
+ steps = response_text.split("\n")
496
+ cleaned_steps = [step.strip() for step in steps if step.strip()]
497
+
498
+ logger.info(f"Extracted {len(cleaned_steps)} action steps")
499
+
500
+ return cleaned_steps
501
+
502
+ except Exception as e:
503
+ logger.error(f"ActionPlanningAgent error: {str(e)}")
504
+ raise
505
+
506
+
507
+ # ============================================================================
508
+ # UTILITY FUNCTIONS
509
+ # ============================================================================
510
+
511
+ def cosine_similarity(vec1: List[float], vec2: List[float]) -> float:
512
+ """
513
+ Calculate cosine similarity between two vectors.
514
+
515
+ Args:
516
+ vec1: First vector.
517
+ vec2: Second vector.
518
+
519
+ Returns:
520
+ Cosine similarity score between 0 and 1.
521
+ """
522
+ vec1 = np.array(vec1)
523
+ vec2 = np.array(vec2)
524
+
525
+ return np.dot(vec1, vec2) / (np.linalg.norm(vec1) * np.linalg.norm(vec2))
requirements.txt ADDED
@@ -0,0 +1,14 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # AI Project Planner - Dependencies
2
+ # For Hugging Face Spaces deployment
3
+
4
+ # Core Framework
5
+ gradio>=4.0.0
6
+
7
+ # OpenAI API
8
+ openai>=1.0.0
9
+
10
+ # Data Processing
11
+ numpy>=1.24.0
12
+
13
+ # Environment Management
14
+ python-dotenv>=1.0.0