---

GRAB action: observe and return a small image

Purpose: make one observation with a single, small window; return the resulting image directly to the client.

Actors: NAOMI or smartflat client; camera server.

Outline: the GRAB action is invoked. The client receives the image resulting from GRAB as a returned argument of the action.

Normal course of events:

1. Client invokes the GRAB action on the server with these arguments:
• Argument1: integration time in seconds.
• Argument2: x-coordinate of lower-left corner of the window.
• Argument3: y-coordinate of lower-left corner of the window.
• Argument4: x-coordinate of upper-right corner of the window.
• Argument5: y-coordinate of upper-right corner of the window.
2. Server validates the arguments.
3. Server clears the detectors, as in the CLEAR action (q.v.).
4. Server exposes the detectors for the given time, as in the INTEGRATE action (q.v.).
5. Server reads out the detectors, as in the READOUT action (q.v.).
6. Server reassembles the regions of the readout into a single rectangle.
7. Server returns good status for the action, with the image sent back in the action's argument list.

• At the time of the action, the camera-server is on-line but has an problem outstanding from the initialisation. Server abandons the normal sequence and does this:
1. Writes an error message directly to the log.
2. Returns status UDAS__NEEDSRESET to the client.
• In step 2, the integration time is negative. Server abandons the normal sequence and does this:
1. Returns an error message to the client.
2. Returns status UDAS__BADEXPTIME to the client.
• In step 2, the window does not fit on the detector. Server abandons the normal sequence and does this:
1. Returns an error message to the client.
2. Returns status UDAS__BADWINDOW to the client.
• In step 3, 4 or 5, the operations on the camera fail. Server abandons the normal sequence and does this:
1. Returns an error message to the client.
2. Returns one of these stati to the client: CAMCMDFAIL, ROFAILED, CAMLINKDOWN.