[TEP007] Isotope stratified file support #764

vg3095 · 2017-07-12T13:42:41Z

This PR is related to adding support for reading stratified abundances with isotopes . This is rebased over PR #762 (Isotope uniform config)

I have added 2 functions -

read_simple_isotope_abundances : To read isotope abundances file in new format
convert_abundances_format : To convert old abundances file format to new one

For sample seeing output:
Notebook
iso_abund.csv
Config

vg3095 · 2017-07-12T15:59:32Z

Travis fail is due to Mac build timeout.

unoebauer

I am wondering whether we can simplify the configuration a bit. In particular, I am thinking of the following changes:

Remove the second line in the stratified file (containing E and Is); it is clear from the header whether the column contains abundances of elements or isotopes; if the identifier in the header contains a number it is an isotope, otherwise an element
remove the additional config identifier filetype: isotopes; we could achieve legacy support by checking for a header line; if it is present we assume the new csv format, if not we assume the old format.

Thoughts on this, @vg3095, @wkerzendorf, @yeganer ?

vg3095 · 2017-07-13T10:20:22Z

I agree on the 1st point, I think we can remove this. I found out that pyne module has a dictionary/mapping between element symbol and its atomic number , and it takes only normal elements( not isotopes). It is found in pyne.nucname.name_zz . If present in this dictionary/mapping, it means its a normal element else isotope.
For second point, I think, if we remove isotopes option, then also , we have to somehow specify that header is present in the config , or I have to check for the first row, if it contains numbers or elements. We can also use try/except inside existing abundances parser. If it can't parse file, then it will re-route to isotope-parser. I don`t know , if this good or not . So, for 2nd point , I am not sure.

wkerzendorf · 2017-07-13T12:20:21Z

tardis/io/model_reader.py

@@ -85,19 +85,66 @@ def read_abundances_file(abundance_filename, abundance_filetype,
    """

    file_parsers = {'simple_ascii': read_simple_ascii_abundances,
-                    'artis': read_simple_ascii_abundances}
+                    'artis': read_simple_ascii_abundances,
+                    'isotopes': read_simple_isotope_abundances}


I would actually call this tardis_model. I think this can be some sort of default model. where you specify things in this way

vg3095 · 2017-07-14T06:21:06Z

I have removed the 2nd header line('E' or 'I') and changed filetype name to tardis_model. Notebook in the description is updated.

unoebauer · 2017-07-14T11:23:55Z

tardis/util.py

+    #Assign header row
+    df.columns = [nucname.name(i)
+                  for i in range(1, df.shape[1] + 1)]


nucname.name(i) will throw an error, if somebody a wrong specifies an element or isotope identifier, right?

Yes, it will throw a Runtime Error with the wrong name of element/isotope. Ex- RuntimeError: Not a Nuclide! XYZ --> 0, where XYZ is wrongly specified element.

unoebauer · 2017-07-14T11:30:13Z

tardis/io/model_reader.py

+            abundance_filename)
+    else:
+        index, abundances = file_parsers[abundance_filetype](
+            abundance_filename)


Can we improve here? Something like

index, abundances, isotope_abundance = file_parser[abundance-filetype]

such that we don't need the if clause?

If, read_simple_ascii_abundances function returns an extra None (instead of (index,abundance), returning index,abundance,None)) , then we can remove this. But I don`t think , its a good idea.

Ok, fine with me

wkerzendorf · 2017-07-14T15:42:25Z

@vg3095 I overlooked that before - but we definitely need tests and documentation for all of this.

vg3095 · 2017-07-17T07:33:25Z

@wkerzendorf I have now updated this PR with tests.

wkerzendorf · 2017-07-17T09:51:07Z

@vg3095 in your examples - you should try to make the abundances add up to one 😉

wkerzendorf · 2017-07-17T09:52:10Z

tardis/io/model_reader.py

@@ -256,3 +264,31 @@ def read_simple_ascii_abundances(fname):
    abundances = pd.DataFrame(data[1:,1:].transpose(), index=np.arange(1, data.shape[1]))

    return index, abundances
+
+
+def read_simple_isotope_abundances(fname, delimiter=','):


the default delimiter should be a whitespace. this is a special option in read_csv.

wkerzendorf · 2017-07-17T09:52:21Z

tardis/io/tests/data/tardis_model_abund.csv

@@ -0,0 +1,11 @@
+C,Mg,Si,Ni56,Ni58


default delimiter should be whitespace.

vg3095 · 2017-07-17T12:07:14Z

@wkerzendorf I have now updated csv files

wkerzendorf · 2017-07-19T08:47:39Z

@vg3095 can you rebase this. @tardis-sn/tardis-core looks good from my side - can someone else have a look?

unoebauer · 2017-07-19T09:20:39Z

tardis/io/model_reader.py

+            abundance_filename)
+    else:
+        index, abundances = file_parsers[abundance_filetype](
+            abundance_filename)


Ok, fine with me

unoebauer · 2017-07-19T09:21:14Z

tardis/util.py

+    #Assign header row
+    df.columns = [nucname.name(i)
+                  for i in range(1, df.shape[1] + 1)]


unoebauer closed this Jul 13, 2017

unoebauer reopened this Jul 13, 2017

unoebauer reviewed Jul 13, 2017

View reviewed changes

wkerzendorf reviewed Jul 13, 2017

View reviewed changes

vg3095 force-pushed the isotope_file_config branch from 9043a33 to b1799e7 Compare July 14, 2017 06:06

unoebauer reviewed Jul 14, 2017

View reviewed changes

vg3095 added 7 commits July 14, 2017 22:04

Added a isotope abund. file parser

518ae3b

Assign isotope abund. in model

a0db95c

Func. to convert old abund. file to new format

bcd7861

Fix typo in comments

5ec16dc

Misc changes in isotope abund. file parser

30d5cfd

Add files related to new isotope config options

23b4001

Add tests for isotope config options

c565565

vg3095 force-pushed the isotope_file_config branch from b1799e7 to c565565 Compare July 17, 2017 03:42

vg3095 added 2 commits July 17, 2017 10:46

Include csv in setup_package.py under io

f36e1b5

Rearrange columns in csv file

58d7a20

wkerzendorf reviewed Jul 17, 2017

View reviewed changes

tardis/io/tests/data/tardis_model_abund.csv Outdated

@@ -0,0 +1,11 @@

C,Mg,Si,Ni56,Ni58

Copy link

Member

wkerzendorf Jul 17, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

default delimiter should be whitespace.

vg3095 added 2 commits July 17, 2017 16:49

Use space delimeter

34a2b0a

Update csv file to make sum of abund. one

fe0283d

unoebauer approved these changes Jul 19, 2017

View reviewed changes

wkerzendorf merged commit 87fe983 into tardis-sn:decay Jul 19, 2017

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[TEP007] Isotope stratified file support #764

[TEP007] Isotope stratified file support #764

vg3095 commented Jul 12, 2017

vg3095 commented Jul 12, 2017

unoebauer left a comment

vg3095 commented Jul 13, 2017 •

edited

Loading

wkerzendorf Jul 13, 2017

vg3095 commented Jul 14, 2017

unoebauer Jul 14, 2017

vg3095 Jul 14, 2017

unoebauer Jul 19, 2017

unoebauer Jul 14, 2017

vg3095 Jul 14, 2017

unoebauer Jul 19, 2017

wkerzendorf commented Jul 14, 2017

vg3095 commented Jul 17, 2017

wkerzendorf commented Jul 17, 2017

wkerzendorf Jul 17, 2017

wkerzendorf Jul 17, 2017

vg3095 commented Jul 17, 2017

wkerzendorf commented Jul 19, 2017

unoebauer Jul 19, 2017

unoebauer Jul 19, 2017

[TEP007] Isotope stratified file support #764

[TEP007] Isotope stratified file support #764

Conversation

vg3095 commented Jul 12, 2017

vg3095 commented Jul 12, 2017

unoebauer left a comment

Choose a reason for hiding this comment

vg3095 commented Jul 13, 2017 • edited Loading

Choose a reason for hiding this comment

vg3095 commented Jul 14, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

wkerzendorf commented Jul 14, 2017

vg3095 commented Jul 17, 2017

wkerzendorf commented Jul 17, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

vg3095 commented Jul 17, 2017

wkerzendorf commented Jul 19, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

vg3095 commented Jul 13, 2017 •

edited

Loading