Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,4 +1,8 @@
|
|
| 1 |
import json
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
import gradio as gr
|
| 3 |
|
| 4 |
from rag.search import search
|
|
@@ -7,12 +11,6 @@ from llm.reasoning import generate_reasoning
|
|
| 7 |
from llm.reasoning import generate_reasoning_from_prompt
|
| 8 |
from rag.ingest import ingest_pdfs_and_web, save_chunks
|
| 9 |
|
| 10 |
-
import base64
|
| 11 |
-
|
| 12 |
-
import os
|
| 13 |
-
|
| 14 |
-
import time
|
| 15 |
-
|
| 16 |
print("🔄 Startar RAG-ingest")
|
| 17 |
DATA_DIR = "rag/data"
|
| 18 |
start_time = time.perf_counter()
|
|
@@ -38,6 +36,16 @@ with open("content.json", encoding="utf-8") as f:
|
|
| 38 |
|
| 39 |
DOC_INDEX = {d["id"]: d for d in DOCUMENTS}
|
| 40 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 41 |
# =====================================================
|
| 42 |
# FUNKTIONER
|
| 43 |
# =====================================================
|
|
@@ -111,15 +119,13 @@ def clear_all():
|
|
| 111 |
|
| 112 |
def format_source_link(chunk: dict) -> str:
|
| 113 |
source = chunk.get("source", "Okänd källa")
|
| 114 |
-
source_type = chunk.get("source_type")
|
| 115 |
|
| 116 |
-
if source_type == "pdf":
|
| 117 |
-
return f"[{source}](/file=
|
| 118 |
|
| 119 |
-
if source_type == "web":
|
| 120 |
return f"[{source}]({source})"
|
| 121 |
|
| 122 |
-
# Fallback – visa åtminstone namnet
|
| 123 |
return source
|
| 124 |
|
| 125 |
|
|
|
|
| 1 |
import json
|
| 2 |
+
import base64
|
| 3 |
+
import os
|
| 4 |
+
import shutil
|
| 5 |
+
import time
|
| 6 |
import gradio as gr
|
| 7 |
|
| 8 |
from rag.search import search
|
|
|
|
| 11 |
from llm.reasoning import generate_reasoning_from_prompt
|
| 12 |
from rag.ingest import ingest_pdfs_and_web, save_chunks
|
| 13 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 14 |
print("🔄 Startar RAG-ingest")
|
| 15 |
DATA_DIR = "rag/data"
|
| 16 |
start_time = time.perf_counter()
|
|
|
|
| 36 |
|
| 37 |
DOC_INDEX = {d["id"]: d for d in DOCUMENTS}
|
| 38 |
|
| 39 |
+
PUBLIC_DIR = "/tmp/gradio/public_pdfs"
|
| 40 |
+
os.makedirs(PUBLIC_DIR, exist_ok=True)
|
| 41 |
+
|
| 42 |
+
for file in os.listdir("rag/files"):
|
| 43 |
+
if file.lower().endswith(".pdf"):
|
| 44 |
+
shutil.copy(
|
| 45 |
+
os.path.join("rag/files", file),
|
| 46 |
+
os.path.join(PUBLIC_DIR, file)
|
| 47 |
+
)
|
| 48 |
+
|
| 49 |
# =====================================================
|
| 50 |
# FUNKTIONER
|
| 51 |
# =====================================================
|
|
|
|
| 119 |
|
| 120 |
def format_source_link(chunk: dict) -> str:
|
| 121 |
source = chunk.get("source", "Okänd källa")
|
|
|
|
| 122 |
|
| 123 |
+
if chunk.get("source_type") == "pdf":
|
| 124 |
+
return f"[{source}](/file={PUBLIC_DIR}/{source})"
|
| 125 |
|
| 126 |
+
if chunk.get("source_type") == "web":
|
| 127 |
return f"[{source}]({source})"
|
| 128 |
|
|
|
|
| 129 |
return source
|
| 130 |
|
| 131 |
|