Section Timings:
Introduction:  planned 30 mins - actual 1:06 pm - 1:24 –                                                               18 mins
How to use Brown Dog Services (1h 30m)

  • 20-30 mins environment setup - actual 1:26 pm - 2:01 pm -                                               25 minutes

Hands-on Details

  • Demonstration of use of BD Fiddle (20 mins) - 2:02 pm - 2:23 pm -                                   21 minutes
  • Conversion Example (20mins):  2:23 pm - 2:52 pm -                                                          29 mins

    BREAK - 8 mins

  • Extraction/Indexing/Retrieval Example (20 mins): 3:00 pm -
  • Conversion & Extraction Example (10 min):  3:39 pm                                                        39 mins ----- 132 mins -- 2 hrs 12 mins ---- 42 mins over

How to add Your Tool to Brown Dog Services (1h 30m)

  • Part 1: Teach to write an extractor (35 mins) - 3:40 pm - 4:12 pm                                      32 mins
    had to stop because of errors

    REVIEW ONLY ------------------------ 4:12 pm -
  • Part 2: Teach to write a converter (35 mins)
  • Part 3: Teach to upload a converter or an extractor to the locally installed Tools Catalog. (20mins)
  • Part 4 (Optional - For advanced user): Dockerize the tool

TOTAL Planned:  3 hrs 30 mins


NOTES/RECOMMENDATIONS:

  • recommendation have 1 screen showing PPT and one showing the actual install (Putty, etc.)
  • possibly have more screen shots in PPT if we can't have a live machine demonstrating (esp windows)
  • screen shots of both linux and windows - leave one screen showing linux ... advance in windows steps ---- possibly have two machines up on screen at same time
  • number the slides for reference
  • send out setup recommendations to have people download and prep their machines ahead of time including signing up for BD
  • need a better way to approve users for BD
  • possibly put it on the usb stick for the conference
  • will need to have team members wandering the room for help during XSEDE conference
  • 3rd slide, the first step, add the move command that is needed: mv (path to downloaded file) ~/.ssh
  • copying and pasting chmod from ppt does not seem to work on MAC - says "file not found"
  • remove "click to add title" from PPT slides for people who download the presentation locally - save PPT as presentation instead of serving the .PPT file
  • prompt name is too large (ubuntu@browndog-xsede-tut:~$) - it takes up too much real estate on the page
  • Linux commands - font is different sizes on the page - runs into the margins and is difficult to see letters
  • BD sign up button broken - http vs https
  • Demonstration of BDFiddle PPT slides - font very small on overhead presentation
  • Click on the GenerateToken,
    it says username but it actually requires an email (confusing,should say email). And if there is an error. It doesn't say error it shows the Key: Token: without any data next to them.
  • URL gets returned with DAP in error (random known issue) - Luigi: That call to polyglot uses a GET and not a POST, like when we uploading a file - I didn't realize and my code only matches POST when rewriting the response for a conversion - Luigi fixed

    BD-1147 - Rewriting of conversion result only works on posting a file and not on url to a file Open

  • ChrisN - could continue just the output does not look like in the slides
    For the JS snippet,is anyone else seeing the message "The bd.js library requires URLs."?
    I don't see the output from the slides
  • RobK - saw this error in the logs
    [Fri Jul 08 14:16:26.426367 2016] [:error] [pid 23154] [client 10.192.106.143:61862] PHP Warning:  file_get_contents(https://bd-api.ncsa.illinois.edu/keys): failed to open stream: HTTP request failed! HTTP/1.1 403 Forbidden\r\n in /var/www/html/bdfiddle/key_token.php on line 24, referer:http://browndog.ncsa.illinois.edu/bdfiddle/key_token.php?host=https://bd-api.ncsa.illinois.edu
  • typos in bd-conversion.py - line 24, line 44, 36 is formats (not format)
  • typos in combined.py - line 23, line 44 (craete), 43 is formats
  • name of files is getting lost in the staple at the top - perhaps staple in middle or use paperclip - recommend tearing it apart???
  • be clear of where they have to enter things in the scripts themselves
  • Conversion slides - Hands-on Exercise - code says path - slide says folder
  • ChrisN - In the Hands-on slide,it's not clear there is more to do than what is listed (you must fill in the "outputs" line and the bd.convert line. I'm not sure if those are intentionally missing from the slide or not.
    I can see people getting hung up on just filling in the token,input_path,output_path and trying to run it.
  • Need 1 zip file of all presentations - so people don't have to download so many files
  • Code that is in handouts needs edited - edited in the PDF? - and highlighted
  • In the extraction example python file, tabs and spaces are mixed and needs fixed
  • You just put in token - you don't need API Key when initializing parameters .... (Marcus's section)
  • In combined PPT - slide 4 - add a slash at the end to the output_path
  • In code - extracted output - most are comments - but there was no comment on the return (line 50) - needs clarification
  • Show where the search term is in the code (Marcus's note)
  • tell them to keep tmux open - open a second terminal
  • nano is not behaving the same as vim - with nano some people are not getting file
  • make sure redis gets into script with full path so it does not have to be started by hand
  • keytoken is giving error because redis/fence was not running because of full path issue
  • xsede is .org not .edu - change in PDFs / handouts
  • VM setup?
  • Make sure what is on a PPT that needs to be read - can be read - font is so small on some of the slides
  • Comment blocks within code so it is more obvious what to change
  • One set of PDF with all the PPTS - One PDF with only what you have to type ... but not 8 PDF separate sets
  • Give stickies to participants - green sticky on top of laptop is OK - pink means need help please come over
  • page with inputs and outputs directories all in one spot - not easy to see/find

 

 

 

  • No labels