File size: 8,461 Bytes
38ab39c 1c47eb5 38ab39c | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 | from fastapi import FastAPI, Form, HTTPException, BackgroundTasks
from fastapi.responses import Response
from cora_engine import CoraEngine
from cora_curator import CoraCurator
from cora_vision import CoraVision
from cora_memory import CoraMemory
import io
import os
import uuid
from pydantic import BaseModel
app = FastAPI(title="Cora API", description="Fake Historical Archive Generator")
engine = CoraEngine()
curator = CoraCurator()
vision = CoraVision()
memory = CoraMemory()
class AgentPrompt(BaseModel):
prompt: str
use_curator: bool = True
@app.get("/health")
def health_check():
"""Checks if the engine and HF connection are ready."""
status = {"status": "online", "model": engine.MODEL_ID}
if not engine.client:
status["status"] = "offline (engine)"
if not curator.client:
status["curator"] = "offline"
else:
status["curator"] = curator.MODEL_ID
# Check Vision/Memory (simple check if initialized)
status["vision"] = "online" if vision.clip_model else "offline"
status["memory"] = "online" if memory.client else "offline"
return status
def archive_generation(image, prompt):
"""Helper to save image and metadata to Visual Memory."""
try:
filename = f"{uuid.uuid4()}.png"
filepath = os.path.join("archive_images", filename)
# Save to disk
image.save(filepath)
# Analyze (Vision)
embedding = vision.embed_image(image)
tags = vision.detect_tags(image)
# Save to Memory (Vector DB)
memory.save(filepath, embedding, prompt, tags)
print(f"✅ Background Archiving Complete: {filepath} with tags {tags}")
except Exception as e:
print(f"❌ Background Archiving Failed: {e}")
@app.post("/agent/generate")
async def agent_generate(request: AgentPrompt, background_tasks: BackgroundTasks):
"""
Agent-friendly endpoint receiving JSON.
Returns the raw PNG image.
"""
try:
# Validate input
if not request.prompt or not request.prompt.strip():
raise HTTPException(
status_code=400,
detail="Prompt cannot be empty. Please provide a description."
)
# 1. Curate (Refine Prompt)
final_prompt = request.prompt
if request.use_curator:
try:
final_prompt = curator.refine_prompt(request.prompt)
except Exception as curator_error:
print(f"Curator failed: {curator_error}, using original prompt")
# Fallback to original if curator fails
final_prompt = request.prompt
# 2. Generate
result = engine.generate_from_text(final_prompt)
# 3. Archive (Background Task)
# We pass a copy or the object itself. Since PIL images are in memory,
# we need to be careful. However, 'result' is a PIL Image.
# It's safer to pass the image object. background_tasks will run after return.
background_tasks.add_task(archive_generation, result, final_prompt)
# Return as PNG
img_byte_arr = io.BytesIO()
result.save(img_byte_arr, format='PNG')
return Response(content=img_byte_arr.getvalue(), media_type="image/png")
except HTTPException:
raise
except ValueError as e:
# User input errors
raise HTTPException(
status_code=400,
detail=f"Invalid request: {str(e)}"
)
except RuntimeError as e:
# Server/API errors
error_msg = str(e).lower()
if "timeout" in error_msg or "took too long" in error_msg:
raise HTTPException(
status_code=500,
detail="Image generation timed out. Try a simpler prompt."
)
else:
raise HTTPException(
status_code=500,
detail=f"Generation failed: {str(e)}"
)
except Exception as e:
print(f"Unexpected server error: {e}")
raise HTTPException(
status_code=500,
detail="An unexpected error occurred. Please try again."
)
@app.post("/v1/archive")
async def generate_archive(
background_tasks: BackgroundTasks,
prompt: str = Form(...)
):
"""
Generates an 'archive' style image from text.
"""
try:
# 1. Curate (Auto-refine for UI)
enhanced_prompt = curator.refine_prompt(prompt)
# 2. Generate
result = engine.generate_from_text(enhanced_prompt)
# 3. Archive (Background Task)
background_tasks.add_task(archive_generation, result, enhanced_prompt)
# Return as PNG
img_byte_arr = io.BytesIO()
result.save(img_byte_arr, format='PNG')
return Response(content=img_byte_arr.getvalue(), media_type="image/png")
except ValueError as e:
raise HTTPException(status_code=400, detail=str(e))
except RuntimeError as e:
raise HTTPException(status_code=500, detail=str(e))
except Exception as e:
print(f"Server Error: {e}")
raise HTTPException(status_code=500, detail="Internal Server Error")
class SearchQuery(BaseModel):
query: str
limit: int = 10
@app.post("/curator/search")
async def curator_search(request: SearchQuery):
"""
Semantic search for the UI gallery with intelligent filtering.
"""
try:
# 1. Embed query
emb = vision.embed_text(request.query)
if not emb:
return {"results": []}
# 2. Extract potential tags from query for filtering
query_lower = request.query.lower()
tag_hints = []
source_hint = None
# Detect cultural/temporal keywords
cultural_markers = {
"roman": ["roman", "rome"],
"greek": ["greek", "greece", "hellenic"],
"egyptian": ["egypt", "egyptian"],
"medieval": ["medieval", "middle ages"],
"renaissance": ["renaissance"],
"enlightment century": ["enlightment century"],
"industrial revolution":["industrial revolution"],
"modern times" : ["modern times", "20th century", "21st century"],
}
for culture, keywords in cultural_markers.items():
if any(kw in query_lower for kw in keywords):
tag_hints.extend(keywords)
# 3. Use hybrid search if we detected cultural markers
if tag_hints:
results = memory.search_hybrid(emb, k=request.limit, tag_filter=tag_hints)
else:
# Fallback to pure semantic if no specific markers
results = memory.search_by_vector(emb, k=request.limit)
# 4. Format result
images = []
if results['ids']:
ids = results['ids'][0]
metadatas = results['metadatas'][0]
distances = results['distances'][0]
for i, uid in enumerate(ids):
path = metadatas[i].get('path')
tags = metadatas[i].get('tags')
prompt = metadatas[i].get('prompt')
if path and os.path.exists(path):
# Convert local path to URL
filename = os.path.basename(path)
image_url = f"http://localhost:8000/archive_images/{filename}"
images.append({
"path": image_url, # Now a URL, not a local path
"tags": tags,
"prompt": prompt,
"score": float(distances[i])
})
return {"results": images}
except Exception as e:
raise HTTPException(status_code=500, detail=str(e))
# Mount static files to serve images to UI if needed
from fastapi.staticfiles import StaticFiles
if not os.path.exists("archive_images"):
os.makedirs("archive_images")
app.mount("/archive_images", StaticFiles(directory="archive_images"), name="archive_images")
if __name__ == "__main__":
import uvicorn
uvicorn.run(app, host="0.0.0.0", port=8000)
|