[diffusion] chore: minor improvements and typo-fixing (#15556)

python/sglang/multimodal_gen/apps/webui/README.md

@@ -1,17 +1,22 @@
 # SGLang Diffusion WebUI User Guide
 
-SGLang Diffusion WebUI provides an intuitive Gradio-based interface for image and video generation, supporting parameter tuning and real-time previews.
+SGLang Diffusion WebUI provides an intuitive Gradio-based interface for image and video generation, supporting parameter
+tuning and real-time previews.
 
 ## Prerequisites
+
 The WebUI runs on Gradio. To get started, install Gradio first:
+
 ```bash
 pip install gradio==6.1.0
 ```
 
 ## Launch WebUI Service
+
 SGLang Diffusion now includes an integrated WebUI. Simply add the `--webui` parameter when starting the service.
 
 ### Launch Text-to-Image Service
+
 ```bash
 SERVER_ARGS=(
   --model-path black-forest-labs/FLUX.1-dev
@@ -20,7 +25,9 @@ SERVER_ARGS=(
 WEBUI_PORT=2333
 sglang serve "${SERVER_ARGS[@]}" --webui --webui-port ${WEBUI_PORT}
 ```
+
 ### Launch Text-to-Video Service
+
 ```bash
 SERVER_ARGS=(
   --model-path Wan-AI/Wan2.2-T2V-A14B-Diffusers
@@ -31,15 +38,18 @@ sglang serve "${SERVER_ARGS[@]}" --webui --webui-port ${WEBUI_PORT}
 ```
 
 ## Port Forwarding
-Once the WebUI service is running, you need to use **SSH port forwarding** to securely access the remote service from your local machine.
 
-In most cases:​ Your IDE (like VS Code, Cursor, etc.) can handle this automatically. Check your IDE's remote development or port forwarding features. Otherwise, execute this command manually.
+Once the WebUI service is running, you need to use **SSH port forwarding** to securely access the remote service from
+your local machine.
+
+In most cases: Your IDE (like VS Code, Cursor, etc.) can handle this automatically. Check your IDE's remote development
+or port forwarding features. Otherwise, execute this command manually.
 
 ```bash
 ssh -L ${WEBUI_PORT}:localhost:${WEBUI_PORT} user_name@machine_name
 ```
-Learn more about port forwarding:​ [Port Forwarding](https://en.wikipedia.org/wiki/Port_forwarding).
 
+Learn more about port forwarding: [Port Forwarding](https://en.wikipedia.org/wiki/Port_forwarding).
 
 ## Interface Instructions
 

python/sglang/multimodal_gen/runtime/entrypoints/diffusion_generator.py

@@ -232,6 +232,7 @@ class DiffGenerator:
                             sample,
                             fps=req.fps,
                             save_output=req.save_output,
+                            # TODO: output file path for req should be determined
                             save_file_path=req.output_file_path(
                                 num_outputs, output_idx
                             ),

python/sglang/multimodal_gen/runtime/entrypoints/utils.py

@@ -91,7 +91,7 @@ def post_process_sample(
                 )
             else:
                 imageio.imwrite(save_file_path, frames[0])
-            logger.info(f"Saved output to {CYAN}{save_file_path}{RESET}")
+            logger.info(f"Output saved to {CYAN}{save_file_path}{RESET}")
         else:
             logger.info(f"No output path provided, output not saved")
 

python/sglang/multimodal_gen/test/run_suite.py

@@ -99,7 +99,7 @@ def run_pytest(files, filter_expr=None):
         print("No files to run.")
         return 0
 
-    base_cmd = [sys.executable, "-m", "pytest", "-s", "-v", "--log-cli-level=INFO"]
+    base_cmd = [sys.executable, "-m", "pytest", "-s", "-v"]
 
     # Add pytest -k filter if provided
     if filter_expr:
@@ -124,20 +124,20 @@ def run_pytest(files, filter_expr=None):
             cmd,
             stdout=subprocess.PIPE,
             stderr=subprocess.STDOUT,
-            text=True,
+            bufsize=0,
-            bufsize=1,
         )
 
-        output_lines = []
+        output_bytes = bytearray()
         while True:
-            line = process.stdout.readline()
+            chunk = process.stdout.read(4096)
-            if not line and process.poll() is not None:
+            if not chunk:
                 break
-            if line:
+            sys.stdout.buffer.write(chunk)
-                sys.stdout.write(line)
+            sys.stdout.buffer.flush()
-                output_lines.append(line)
+            output_bytes.extend(chunk)
 
-        returncode = process.poll()
+        process.wait()
+        returncode = process.returncode
 
         if returncode == 0:
             return 0
@@ -152,7 +152,7 @@ def run_pytest(files, filter_expr=None):
             return 0
 
         # check if the failure is due to an assertion in test_server_utils.py
-        full_output = "".join(output_lines)
+        full_output = output_bytes.decode("utf-8", errors="replace")
         is_perf_assertion = (
             "multimodal_gen/test/server/test_server_utils.py" in full_output
             and "AssertionError" in full_output

python/sglang/multimodal_gen/test/server/test_server_utils.py

@@ -322,6 +322,7 @@ class ServerManager:
                     with pipe:
                         for line in iter(pipe.readline, ""):
                             sys.stdout.write(line)
+                            sys.stdout.flush()
                             file.write(line)
                             file.flush()
                 except Exception as e:
@@ -360,6 +361,8 @@ class ServerManager:
         """Wait for server to become ready."""
         start = time.time()
         ready_message = "Application startup complete."
+        log_period = 30
+        prev_log_period_count = 0
 
         while time.time() - start < self.wait_deadline:
             if process.poll() is not None:
@@ -378,8 +381,10 @@ class ServerManager:
                     logger.debug("Could not read log yet: %s", e)
 
             elapsed = int(time.time() - start)
+            if (elapsed // log_period) > prev_log_period_count:
+                prev_log_period_count = elapsed // log_period
                 logger.info("[server-test] Waiting for server... elapsed=%ss", elapsed)
-            time.sleep(5)
+            time.sleep(1)
 
         tail = self._get_log_tail(stdout_path)
         raise TimeoutError(f"Server not ready within {self.wait_deadline}s.\n{tail}")