Difference between revisions of "CalCOFI Drifter Realtime Webpage Info"
Kirk ireson (talk | contribs) |
Kirk ireson (talk | contribs) |
||
| Line 180: | Line 180: | ||
<br> | <br> | ||
<b>NOTES:</b><br> | <b>NOTES:</b><br> | ||
| − | -Total bit count of all 7 lines must equal the checksum value or | + | -Total bit count (a very non trivial calculation) of all 7 lines must equal the checksum value or |
message block is considered garbled and thus ignored.<br> | message block is considered garbled and thus ignored.<br> | ||
<br> | <br> | ||
Revision as of 13:03, 26 June 2009
Contents
CalCOFI SVP Drifter Realtime Webpage Wiki
NOTES AND THINGS TO KNOW ABOUT THE SVP REALTIME PAGE
Jun, 2009 Kirk
Webpage: SVP Realtime Page
There are three main parts needed to display the data (details below):
1. Fetching the data (using Python script running on dub-locl)
2. Parsing the data (using PHP script running on website)
3. Displaying the data (using Javascript running on your browser)
1. FETCHING THE DATA
On dub-locl.icess.ucsb.edu in /localhome/kirk/scripts/ there is a script 'expect_script_to_fetch_argos_data.py'. It uses 'pexpect', a Python version of the 'expect' scripts, which allows you to script what would normally be an interactive terminal session. This is necessary in this case because we have to obtain the data via a telnet session.
- There is essentially nothing to configure it, and it's well documented should you need to change anything. It is set to fetch the last 24hours worth of positions.
- It may need to be updated with drifter numbers as more are sent drifting.
- It is set to run every 6 hours via a "cron job" in the kirk account. To edit this job, as kirk, at the Terminal, type 'crontab -e' and use 'vi' style editing commands, or 'crontab -l' to just show it. (To become kirk, login as drift, then in a Terminal, type 'sudo su -' (then type the drift password) to become root, then 'su kirk' to become kirk.)
- It saves the output of this session to /home/locl/file_sharing/SVP_CALCOFI_LATEST_DATA.txt on the ICESS network (which is accessed on dub-locl via /mnt/locl/, which should be mounted automatically if the system reboots, see /etc/fstab)
2. PARSING THE DATA
When someone visits the webpage (http://www.icess.ucsb.edu/drifter/realtime-SVP/), in the ajax.js file there is instruction to fetch the data. This 'fetch' is accomplished by calling 'parse_argos_data.php' in that directory. This php script reads the file saved above (in /home/locl/), ignores all the non-data lines, and processes/cleans up the crazy ARGOS format output from the telnet session.
It outputs 1 record per position report in this format:
drift_no,YY-MM-DD,HH:MM:SS,milliseconds_since_Jan01_1970,orig_report_num,lat,lon,position_type(1=GPS,0=ARGOS),SST(C),batt(V),base_time,record_num
e.g.
75428,2007-04-03,10:33:37,1175610817000,137,+26.6979,-081.9557,1,NaN,12.4,2007-04-03_11:33:30,2 (a gps position)
- There is very little configuring it, and it's well documented should you need to change anything.
- Its output can be seen easily: http://www.icess.ucsb.edu/drifter/realtime-SVP/parse_argos_data.php
3. DISPLAYING THE DATA (at realtime page)
The Google mapping code was adapted from the Iridium/Microstar realtime page at http://www.icess.ucsb.edu/drifter/realtime/
The two main javascript files are ajax.js and map.js.
The javascript files are compressed using http://www.icess.ucsb.edu/drifter/realtime/JavascriptCompressor.html, so changes made to the xxx__UNCOMPRESSED__.js files won't show up on the page unless they are compressed and saved without the "__UNCOMPRESSED__" portion in the name, OR at the beginning of index.php those two types of files are commented/uncommented out as appropriate.
POTENTIAL PROBLEMS / TROUBLESHOOTING
Prob: The data file (/home/locl/file_sharing/SVP_CALCOFI_LATEST_DATA.txt) is not getting saved/updated
- Is the dub-locl computer up?
- Is /mnt/locl mounted on dub-locl? It should come up automatically
- Has 6 hours passed since the last time it ran? It only runs every 6 hours
- Try fetching the data yourself via telnet (do these commands at a terminal prompt):
telnet datadist.argosinc.com hansen (at the 'Username:' prompt) aomlus (at the 'Password:' prompt) prv,7325,ds,,93561-93587 (at the '/' prompt, to get the previous 9 day's worth of data for drifter #93561-93587 ) lo (at the '/' prompt, to log out)
NOTE: if you want to get data from a specific time period, use julian dates and GMT hours after the comma after 'ds'. For example:
prv,7325,ds,174/12-175/10,93561-93587 (for June 23rd at 12Z through June 24th at 10Z)
Prob: The data file (/home/locl/file_sharing/SVP_CALCOFI_LATEST_DATA.txt) shows no data
- There probably aren't any drifters talking in the last 24hours (look for the line "No data available or authorization failure" in the file or try manually getting the data)
INFO ABOUT TIME CALCULATION
Basically each data block is for three hourly reports, and data blocks can be repeated (I
have no idea how that's determined), and there is a slight adjustment
to the time based on "fix delay" (or "Length of Acquisition", which is
usually on the order of <60 seconds).
For example (showing the second to last block in the existing file as
I write this):
07325 93575 15 27 N 3 2009-06-25 20:18:59 34.416 240.156 0.000 401652653
2009-06-25 20:21:22 3 12 62 678 94
00 01 344156 01
1198452 01 04 689
01 344155 01 1198448
01 04 661 01
344156 01 1198448 01
04 587 00
This *first line* contains ARGOS only info:
prog_num 07325 drift_num 93575 lines_in_sat_pass 15 num_sensors 27 sat_name N resolution 3 time 2009-06-25 20:18:59 xmit_lat 34.416 xmit_lon 240.156 xmit_alt 0.000 xmit_freq 401652653
The next 7 lines contain 3, hourly GPS reports.
Line 1: (2009-06-25 20:21:22 3 12 62 678 94)
BASE Date/Time 2009-06-25 20:21:22 # identical messages rcvd 3 Drogue count over last half hour 12 Battery voltage = n * 0.2 62 -> 12.4V SST(C) = x * 0.043 - 5 678 -> 24.154 Checksum 94
Line 2: (00 01 344156 01)
Msg ID 00 -> this is 1st message sent Valid Pos #1 (1=good, 0=bad) 01 -> good Lat #1 (ten-thousandths of degree) 344156 -> 34.4156 Lat #1 Ref (0 = South, 1 = North) 01 -> North
Line 3: (1198452 01 04 689)
Lon #1 (ten-thousandths of degree) 1198452 -> 119.8452 Lon #1 Ref (0 = East, 1 = West) 01 -> West Pos #1 Length of Acq (seconds = n * 4) 04 -> 16(seconds) SST #1 (C) = x * 0.043 - 5 689 -> 24.627
Line 4: (01 344155 01 1198448)
Valid Pos #2 (1=good, 0=bad) 01 -> good Lat #2 (ten-thousandths of degree) 344155 -> 34.4155 Lat #2 Ref (0 = South, 1 = North) 01 -> North Lon #2 (ten-thousandths of degree) 1198448 -> 119.8448
Line 5: (01 04 661 01)
Lon #2 Ref (0 = East, 1 = West) 01 -> West Pos #2 Length of Acq (seconds = n * 4) 04 -> 16(seconds) SST #2 (C) = x * 0.043 - 5 661 -> 23.423 Valid Pos #3 (1=good, 0=bad) 01 -> good
Line 6: (344156 01 1198448 01)
Lat #3 (ten-thousandths of degree) 344156 -> 34.4156 Lat #3 Ref (0 = South, 1 = North) 01 -> North Lon #3 (ten-thousandths of degree) 1198448 -> 119.8448 Lon #3 Ref (0 = East, 1 = West) 01 -> West
Line 7: (04 587 00)
Pos #3 Length of Acq (seconds = n * 4) 04 -> 16(seconds) SST #3 (C) = x * 0.043 - 5 587 -> 20.241 Null bits (only for padding, ignore)
NOTES:
-Total bit count (a very non trivial calculation) of all 7 lines must equal the checksum value or
message block is considered garbled and thus ignored.
-Times of GPS reports are calculated thusly:
1) Take BASE time, subtract (Msg ID*3) [that is, if message has been
repeated, it is at least 3 hours old -- I have a switch in my code
(it's turned on for now) which will ignore messages with a message
ID>0 because the original may be also present, and it's possible the
duplicate isn't an exact duplicate and then you get slightly different
reports at similar times]
2) Then, subtract 2 hours for Pos #3, and 1 hour for Pos #2 and 0 for Pos #1
3) Then subtract the associated small "Length of Acq"
Thus, the 2nd GPS time (i.e. #2) would be calculated thusly:
BASE time: 2009-06-25 20:21:22 minus 0 hours (because Msg ID is 0) minus 1 hour (because it's Pos #2) minus (pos #2 length of acq) 16 seconds equals: 2009-06-25 19:21:06
An additional wacky thing is that due to repetition of the messages,
or god knows what, the calculations of the reports you see aren't
necessarily in chronological order. So, after doing all the
calculations, I then sort them chronologically before outputting
(within the parse_argos_data.php script), so that when showing the
track lines, the points are in order.
