# Unpredictable behavior of paraview-lite
**URL:** https://discourse.paraview.org/t/unpredictable-behavior-of-paraview-lite/13067
**Category:** ParaView Support
**Created:** [October 20, 2023, 6:20am UTC](https://discourse.paraview.org/t/unpredictable-behavior-of-paraview-lite/13067 "2023-10-20T06:20:34Z")
**Posts on this page:** 4
**Page:** 1
### Author:  [@fashing](https://discourse.paraview.org/u/fashing)
#### Post date: [October 20, 2023, 6:20am UTC](https://discourse.paraview.org/t/unpredictable-behavior-of-paraview-lite/13067/1 "2023-10-20T06:20:34Z")
Hi Paraview,
I would like to share my weird experience since I have no clue about it.
I have some volume data that stored the signed distance value and would be visualized by the `contour` filter. Some of the arrays have opening surfaces in paraview, and I thought it is because somehow my data is broken.
One day, I launched paraview-lite and opened the `show annotation` parameter in `rendering view settings` to track the fps. I loaded the data and contoured them, then I found it takes a lot of time to set a proxy invisible (before I show the annotation, it takes no time to set a proxy invisible). I reload the browser while it is still progressing (the blue progress bar on top is running). And I got these error messages:
```auto
ERROR:asyncio:Task exception was never retrieved
future: exception=AssertionError()>
Traceback (most recent call last):
File "C:\Program Files\ParaView 5.11.2\bin\Lib\site-packages\wslink\backends\aiohttp\ __init__.py", line 638, in sendWrappedMessage
await ws.send_bytes(attachments[key])
File "C:\Program Files\ParaView 5.11.2\bin\Lib\site-packages\aiohttp\web_ws.py", line 307, in send_bytes
await self._writer.send(data, binary=True, compress=compress)
File "C:\Program Files\ParaView 5.11.2\bin\Lib\site-packages\aiohttp\http_websocket.py", line 685, in send
await self._send_frame(message, WSMsgType.BINARY, compress)
File "C:\Program Files\ParaView 5.11.2\bin\Lib\site-packages\aiohttp\http_websocket.py", line 656, in _send_frame
await self.protocol._drain_helper()
File "C:\Program Files\ParaView 5.11.2\bin\Lib\site-packages\aiohttp\base_protocol.py", line 84, in _drain_helper
assert waiter is None or waiter.cancelled()
```
And the strange thing happened here. When the browser is reconnected, the opening surface of the counter got close and became the correct image I expect. Setting a proxy to invisible also becomes costing no time.
I think there are two mysteries here:
1. Showing annotation make modification on proxy become time costly.
2. Unexpectedly reloading (while processing) somehow “fix” the image.
Anyone having any clue about it would be very helpful to me. Thank you!
---
### Author:  [@mwestphal](https://discourse.paraview.org/u/mwestphal)
#### Post date: [October 23, 2023, 7:41am UTC](https://discourse.paraview.org/t/unpredictable-behavior-of-paraview-lite/13067/2 "2023-10-23T07:41:14Z")
please share precise stesp to reproduce.
---
### Author:  [@fashing](https://discourse.paraview.org/u/fashing)
#### Post date: [October 23, 2023, 8:24am UTC](https://discourse.paraview.org/t/unpredictable-behavior-of-paraview-lite/13067/3 "2023-10-23T08:24:28Z")
Hi Mathieu,
Sorry that I cannot provide the file I used. However, I found something today and the issue is clearer now.
The root cause is that the contour result sometimes differs in ParaView and ParaView-Lite. e.g., the Bounds of contour in Paraview is (a, b, c, d, e, f) and (a, b, c, d+1, e, f) in ParaView-Lite.
I am now looking into how contour is called in ParaView-Lite. If you have any insight about how this happens, I’ll be happy to hear so!
---
### Author:  [@fashing](https://discourse.paraview.org/u/fashing)
#### Post date: [October 24, 2023, 6:01am UTC](https://discourse.paraview.org/t/unpredictable-behavior-of-paraview-lite/13067/4 "2023-10-24T06:01:02Z")
Hi @mwestphal,
I found the root cause today.
I have a `Partitioned Dataset Collection` whose `Bounds` are `(0, 200, 0, 200, 0, 600)`.
And one of the `Partitioned Dataset` has `Bounds` `(0, 200, 0, 199, 200, 400)`, and its contour has `Bounds` `(0, 200, 0, 199, 200, 400)`.
The unexpected shrink of the `Partitioned Dataset` on y-axis leads to the open surface of the contour filter.
If I check on `Merge Image Volumes`, the `Bounds` of the contour becomes `(0, 200, 0, 200, 200, 400)` and hence the open surface closes. I accidentally check the merge volume option in ParaView-Lite, that’s why the contour only closes there.
Do you have any idea about the unexpected shrink on `Partitioned Dataset`? Thank you!