pheelcool / dsm2-vista Goto Github PK

This will be a new feature. The idea is simply to combine the UI (vista) and 
scripting (vscript). 
There are a couple of concepts here. One is to give a vscript shell, 
available at any time from the UI and has a few predefined variables such as 
the current view, selected data in a view etc. The other is to have the 
ability to run scriptlets, essentially scripts meant to be run within the UI 
as they will reference those predefined variables.

One scenario that should be addressed by this feature is being able to take 
a selected data table and run a criteria to select a few rows from that 
table view and then set those selected rows value (or flag values) to 
desired constant or function. 

1. Accumulate a dss record or an array over a time period.
2. Linear interpolation of a dss record or an array.
3. Lookup function with an ability to interpolate if the value looking up falls 
between two values in the lookup array.

Thanks,
Chandra

create reference ref with 15min regular time series data. calling per_avg(ref) 
on this reference changes this original reference 'ref' to 1day. That should 
not be so, the 15min original reference should stay at that and the return 
reference should be 1day

Hi,
I have 2 server with two application using dsm2-vista accessing the same DSS. 
One server write the DSS, the other one read DSS. When happens that the servers 
operates in the same time (one reading, the others writing) a JNI exception is 
raised and the JVM of one of the two go down.
Can you help or clarify me about this please.
Thanks.

Currently ui does not indicate at what point user is zoomed in/out. It does not 
allow paging by default in zoomed state

Indicate zoom in/out by providing scroll bars and/or overview window with 
rectangle that can be moved 

Add ability to shift time series values in time in units of minutes

Irregular to Regular using snap or sampling.
snap has options of nearest, backwards, forwards
sample has options of last value in time, linear interpolation

importing vtidefile fails with a message that hdf5 related class is missing. 
The workaround is to add the hdf5 jars to the batch file used to launch 
vscript.

Water Data Library (http://www.water.ca.gov/waterdatalibrary/) is a source of 
grab sample data. A script is needed to query the stations within a map bounds 
(rectangle) and query for stations within that. Then it should loop over each 
station id and capture the data as irregular time series and write it out a dss 
file.

Currently the pathnames cannot be edited from the GroupTable View. Allow in 
place editing of these values. 

Changing pathnames does not currently change the pathname for that dataset in 
the dss file, so allow this feature only for datasets that have been generated 
as a result of operations on exiting data sets.

Hi,
I'm JAVA/C++ developer, and I need to show a DSS archive via web application. 
To do this I'm useing the Smartclient framework (http://www.smartclient.com/) 
to make interfaces (layout, catalog list, path list of cataolg and javascript 
interaction as filter, sorting, selecting), amcharts (http://www.amcharts.com) 
to dynamic load and create interactive graphs, and some part of your works 
(dsm2-vista) to reads and query DSS files. I'm actually testing and reviewing 
the project code to create a structured software. If you are interested I can 
share my works, for your and others user benefit. What do you think? 

Get a data reference from a tidefile. 
set a time window on it which is smaller than the timewindow available
get its data set and that data set's timewindow 
compare the reference timewindow and data set time window and see the 
mismatch

This functionality exists but perhaps does not allow for non overlapping time 
windows. It should do so.

Currently flags are set based on user selection of a time interval. This 
functionality will allow the selection to be enhanced to be a rectangular box 
selecting both a time interval and a range of values within that time interval. 

The user will click on a beginning point and then move the mouse. this will 
show the box within which the curve will be acted upon. The user then gets 
dialogs similar to flag setting which will allow the user to set the flags.

I tried to make the data query by using findparts function in vscript, a
null set was returned when there were special characters in the string. An
example is shown as below. 

>>> g1 = findparts(g,b='SAC_YOLO',c='FLOW')
>>> ref1=g1[0]
>>> g1 = findparts(g,b='SAC+YOLO',c='FLOW')
>>> ref1=g1[0]
Traceback (most recent call last):
  File "<console>", line 1, in <module>
TypeError: 'NoneType' object is unsubscriptable
>>>

Purpose of code changes on this branch:
Instead of maintaining java bindings to the heclib use the heclib bindings 
provided by HEC in DSSVue

When reviewing my code changes, please focus on:
Changes made to address issue 6 as well as the switching of the code

After the review, I'll merge this branch into:
/trunk

Create a easy conversion for Time to Date as HEC Time uses a different epoch 
date than Java

In observed data there can be a fair amount of noise. Typically this is removed 
using manual identification of outliers. 
Add a few simple functions to assist in speeding this up.
a.) where(timeseries, conditional) creates a timeseries of 0 and 1 where 
conditional is false (0) and true (1)
b.) apply(timeseries, conditional, [apply_function]) clears the value to 
missing where conditional is false and keeps it where conditional is true. Also 
optionally have an apply_function which gets the time series value at that 
point and the result of the conditional and changes the value for the time 
series. A funtion here can flag the data to desired flag value as well.
c.) shift(timeseries, shift) creates a shifted time series with the shift in 
time units of minutes
d.) auto_difference(timeseries) creates a time series replacing x(t) with 
x(t)-x(t-1). Its equivalent to timeseries - 
shift(timeseries,timeseries_interval_in_minutes)

Creating a reference with a given timewindow from DataReference from 
python is not very pythonesque. add a __getslice__ method on DataReference 
very similar to RegularTimeSeries

The DSS pathname consists of A-F parts and is represented by a String. This 
string is parsed in java to a Pathname object, however manipulating it (e.g. 
setting/getting a path part) is cumbersome.
Add methods to make it seem that Pathname is a associative array of path 
parts referenced by "A"-"F" (case insensitive pls)

A busy cursor should be displayed when opening a large DSS file. This is not 
happening and the UI becomes unresponsive during long operations.

The math operations panel in the current gui is cumbersome where it operates on 
current selection. Enhance the panel by adding an expression box where users 
can refer to a time series using its current row and writing a mathematical 
expression consisting of +,-,*,/ operators as well as math functions such as 
per_avg, mov_avg, godin etcetra.

Change attribute of a data set from PER-AVER to INST-VAL from the table option 
of editing attributes does not work.

Given datasets (say regular timeseries), merge, moving average, period 
average should not require creating pseudo references just to get those 
operations done. These should be available on the datasets.

Purpose of code changes on this branch:


When reviewing my code changes, please focus on:


After the review, I'll merge this branch into:
/trunk/vista

fail to add flags in the DSS file \\Delta-mod\Ydrive\Observed
Data\CDEC\CDEC-its-auto.dss

This happens intermittently and unpredicatbly. The JVM quits with a access 
violation error with the stack trace pointing to DSS.dll. 

There probably is a memory leak in DSS.dll and this needs to be resolved by 
either switching to the dll provided in HecDssVue or addressing it in the 
wrapper code.

try using ds != None. It'll fail as it calls the overloaded operator method in 
the dataset or derived types and that does not handle this case correctly

Tidefile format changed with v8.0b6. and vista can't read the data from the 
hdf5 anymore

For different version of VISTA, there is naming inconsistency while
producing period averaged time series. For example, the latest version
names the series as "SAC_PERAVG" while the old version names the series as
"SACPER-AVG". 
This may cause problems while using vscript to process those series.
However, this is a minor issue since users can fix by considering those two
conditions in their scripts. 

Hi,
I have a project that use both dsm2-vista and HEC-HMS-3.4, but I have some 
problem at runtume (method not found exception) because the same class 
Heclib.class have different methods in the same heclib.jar present in 
dsm2-vista and HMS-3.4 and I need both in classpath.

Thanks.

open a dss file, close the second frame (listing data paths). try to open again 
by using Group|Open from the first frame menu. This does not work

Currently the flag editing selection only selects a range of time values. 
Enhance the selection to consist of both range of time values and range of 
values of the curve

Additionally add the ability to edit the values (not just flags) using drag and 
drop and/or an entry box for new values and/or shifting the values by a set 
amount

Purpose of code changes on this branch:
 - Cleaning up of scripts that wrap the functionality in Java. 
   * Naming conventions (use standard python conventions)
 - Document functions that are not straightforward
 - Refactoring suggestions to avoid code duplication

What steps will reproduce the problem?
1. go to vista and make 1 YEAR period average of any time series
2. error occurs when trying to view the data as either table or graph
3. "java.lang.IllegalArgumentException: All references to be graphed are null"