Happy Christmas all!
What with the holidays, end of ear etc its been a bit busy.
its absolutely great to see CLS in quicklisp!
However, Tony, I have pushed my recent changes - warning only lightly tested
Things of note
1. Dataframe is now spread across different files for ease of management
2. query stuff is in dataframe-query, though after thinking about things
I have a some better ideas , mainly stolen from R
3. have an import function for fixed format data files now, this
introduced a dependancy on data-format-validation a libraray that I have
used in the past and so far, has been pretty solid. Not wedded to it,
but its convenient and flexible. Its in import.lisp,
file.fixed->dataframe. There is a sample fixed frormat file in the data
directory
4. dataframe metadata - there is a new slot called 'variables' and this
is where I stash away things like column types, print widths and
probably will add other things like a list of unique entries for
categories etc. At the moment the scheme does create overhead when we
create a dataframe, but for anything under 10000 rows on my macbook I
have not seen an issue. Performance testing soon when I start crunching
my weather station records (several million rows.)
5. There is a utility function called dfgrid which will create a grid
to pass to gsl. Not tested as yet, but thats the main short term vehicle
for accessing GSLL viz
(LINEAR-FIT (dfgrid df :x) (dfgrid df :y))
6. I also have a antik:grid dataframe experiment too, which kind of
worked. My conern is now copying large amounts of data , so some thought
has to be given as to exactly what stays in CLS and what gets outsorces
to GSLL. I need the probability distrubtions back so will produce a
mapping to gsll for that soon.
7. i am going to stop working on dataframes for a short while and get
the gnuplot interface going, nothing fancy, but enough to see
scatterplots, trends, errorbars that sort of thing. We can look at the
G-O-G approach once thats done.
Tony, ping me when you have looked through the changes - its time for me
to retire this local repo and get in synch with the mainstream.
Happy New Year all!
> state up, I can work on integration. --
> You received this message because you are subscribed to the Google
> Groups "Common Lisp Statistics" group.
> To post to this group, send email to
lisp...@googlegroups.com.
> To unsubscribe from this group, send email to
>
lisp-stat+...@googlegroups.com.
> Visit this group at
http://groups.google.com/group/lisp-stat?hl=en.
>
>