Docs: self-hosting guidance + FastAPI example

README.md

@@ -161,6 +161,10 @@ CHATGPT_API_KEY=your_openai_key_here
 python3 run_pageindex.py --pdf_path /path/to/your/document.pdf
 ```
 
+### Self Hosting
+
+This repository does not ship a production API server. For a minimal self-hosted HTTP wrapper, see `SELF_HOSTING.md`.
+
 <details>
 <summary><strong>Optional parameters</strong></summary>
 <br>

SELF_HOSTING.md

@@ -0,0 +1,42 @@
+# Self Hosting (Open-Source)
+
+This repository focuses on **generating PageIndex tree structures** locally (PDF → tree JSON). It does **not** ship the same hosted SaaS/API backend as `api.pageindex.ai`.
+
+If you want to “self host” for internal use, the typical approach is:
+
+1) run the CLI (`run_pageindex.py`) in your own environment, or  
+2) wrap the CLI/library with a small HTTP server.
+
+## Minimal FastAPI Server
+
+An example server is provided at `examples/self_host_fastapi.py`. It exposes a single endpoint that accepts a PDF upload and returns the generated tree JSON.
+
+### Install
+
+```bash
+pip3 install -r requirements.txt fastapi uvicorn
+```
+
+### Configure
+
+Create a `.env` in the repo root with your LLM settings (example):
+
+```bash
+CHATGPT_API_KEY=
+CHATGPT_MODEL=gpt-4o-2024-11-20
+CHATGPT_BASE_URL=https://api.openai.com/v1
+```
+
+### Run
+
+```bash
+uvicorn examples.self_host_fastapi:app --reload --port 8000
+```
+
+### Call
+
+```bash
+curl -F "file=@/path/to/document.pdf" "http://localhost:8000/index"
+```
+
+Addresses #82

examples/self_host_fastapi.py

@@ -0,0 +1,58 @@
+import os
+import tempfile
+
+from fastapi import FastAPI, File, HTTPException, UploadFile
+
+from pageindex.page_index import page_index_main
+from pageindex.utils import config
+
+
+app = FastAPI(title="PageIndex (Self-Hosted Example)")
+
+
+@app.get("/health")
+def health():
+    return {"status": "ok"}
+
+
+@app.post("/index")
+def index_pdf(
+    file: UploadFile = File(...),
+    model: str = "gpt-4o-2024-11-20",
+    toc_check_pages: int = 20,
+    max_pages_per_node: int = 10,
+    max_tokens_per_node: int = 20000,
+    if_add_node_id: str = "yes",
+    if_add_node_summary: str = "yes",
+    if_add_doc_description: str = "no",
+    if_add_node_text: str = "no",
+):
+    filename = file.filename or ""
+    if not filename.lower().endswith(".pdf"):
+        raise HTTPException(status_code=400, detail="Only .pdf uploads are supported")
+
+    pdf_bytes = file.file.read()
+    if not pdf_bytes:
+        raise HTTPException(status_code=400, detail="Empty upload")
+
+    tmp_path = None
+    try:
+        with tempfile.NamedTemporaryFile(delete=False, suffix=".pdf") as tmp:
+            tmp.write(pdf_bytes)
+            tmp_path = tmp.name
+
+        opt = config(
+            model=model,
+            toc_check_page_num=toc_check_pages,
+            max_page_num_each_node=max_pages_per_node,
+            max_token_num_each_node=max_tokens_per_node,
+            if_add_node_id=if_add_node_id,
+            if_add_node_summary=if_add_node_summary,
+            if_add_doc_description=if_add_doc_description,
+            if_add_node_text=if_add_node_text,
+        )
+
+        return page_index_main(tmp_path, opt)
+    finally:
+        if tmp_path and os.path.exists(tmp_path):
+            os.remove(tmp_path)