Talk:Implement ALRS
From PTAGISWiki
Carters 15:02, 27 July 2006 (PDT)
Dave is investigating how to maintain existing functionality of RKM and Valid_Tbl and provide a batch update of ALRS (P3 and FDVL requirements).
There is a 2 phase process for the update of the approximately 750 locations in PTAGIS:
- Incorporate new location definitions into ALRS.
- Insure that subsequent entries are incorporated properly.
Dave received an LLIDStream table from Van Hare. He will merge this data with PTAGIS location information.
- Need to add LocationID column to ALRS schema Doug?
- Need to add modification date to ALRS application (perhaps a transaction log?) Doug?
Ryan's workflow analysis
Rday 13:54, 2 August 2006 (PDT)
In order to make the ALRS data entry reasonably simple for end users, there needs to be a wizard for each common data entry use case. Over the past few months, the most common use case has been the addition of a new Tagging, Release, and Recovery site. The workflow for this use case currently looks like this: Add ALRS site.
The number of use cases is not that great, but it is complicated by the need to determine the existence of dependent objects like river segments before adding an object to ALRS.
Carters 13:35, 4 August 2006 (PDT) Meeting with Ryan
Ryan will continue his workflow analysis on adding new sites. Need to:
- Develop the top 5 use cases
- Design a work flow for each use case
It's reasonable to do these by December, 2006.
Carters 15:55, 31 August 2006 (PDT) Meeting with Dave
- Dave has developed lists of unique tag, release sites from tag_hdr, recapture sites and re-release sites from recap_hdr and sites from mort_hdr. In addition he has gathered the list of hatchery sites. This is what we have right now.
- Question for Doug: What are the implications of using more detailed (high granularity) and less abstract categories in the activities table? We have hatchery locations not associated with any PIT tagging events. How would these be incorporated into the model and how do we associate an activity with it?
- Another question: can or should a site be both a hatchery and an acclimation pond? It is easier to locate and maintain records if there is a single entry.
- Hopefully we could have all our data ready to go by end of September as long as the trap data is validated by the contributors.
Rday 09:36, 12 September 2006 (PDT) Model for public interface to ALRS
Even before finishing the analysis of the common use cases, I think I have a way to accomodate them in one screen.
Many of the use cases are similar, but differ when there is a missing dependency for a site or activity that the user wants to add. The dependency structure looks like this:
- Activity depends on an existing location
- Location depends on an existing river segment
- River segment depends on an existing HUC
I'll ignore the case of needing to add a HUC for this discussion, since they are so stable.
If the entry form for adding a new activity includes dependent sections for location and river segment, the form could be made to handle most use cases. A rough layout of the form is here: http://snapper.psmfc.org/ptagisajax/www/com.ptagis.ptagisajax/ptagisajax.html
Carters 12:10, 13 September 2006 (PDT) Meeting with Dave
Carter and Dave walked through the ALRS schema and discussed issues related to how to populate the schema. Dave illustrated the location_matrix which can be used as the main source for populating the activity_location table. Dave will continue to work on this. Targeting completion for October.
Dave will provide input to Ryan on user interface design for the 'add a new location' application.
We discussed drilling into the activity portion of this model more after the basic location and activity location / river segments were installed. This would involve potential schema modifications to define activities in a more granular form, incorporate flag codes, spec doc updates and new u/i requirements for P4.
Rday 11:15, 24 October 2006 (PDT) web form feedback
Lat/Lon should only be required for traps Streamnet ID is optional
DWR Direct web remoting if you are writing javascript code
Some of the picklists should be populated from streamnet. But the data sources need some smart filtering because some of those lists have 10,000 entries.
The data should be entered by our users, but marked as provisional until approved by PTAGIS staff.
Do we really need to ask for start and stop dates? Only for modifying traps. Prompt for dates in the PTAGIS admin interface.
Rday 10:26, 30 October 2006 (PST) Next priority is to get GWT to talk to database
Collecting resources to figure out how to use the GWT framework to access existing DAO methods. Doug has documented Generic Reference-Data that I'll be interfacing with. Now I need to pick a strategy to link the code running in the user's browser to those DAOs. It may mean that I have to construct a proxy web service that will accept requests for data from browsers and hand them off to Doug's DAOs and then returning the results to the web clients.
I'm collecting a page of resources that are helpful in building this linkage here: Google web toolkit.
