Write ALARA output to SQLite by anu1217 · Pull Request #75 · svalinn/activationDB

anu1217 · 2026-01-20T05:10:56Z

This PR introduces a structure to the sqlite activation results database and writes some preliminary results to it.

The run_lbl column is used to keep track of the ALARA runs from which results were produced (using the number of pulses, duty cycle %, and active burn time).

The block_name column has the effect of tracking the parent nuclide. The (flattened) irradiation time in seconds, flux spectrum shape, and number density all have separate columns.

This database should be expanded to include the average flux magnitude (pending #14)

fixes #35

gonuke

I started reviewing this a while ago and never got around to finishing it.

Here are some specific suggestions.

Overall, I think this PR is mixing things based on what your task was at the time and should maybe be broken down somehwat, aka separation of concerns.

Fundamentally, if you are going to rely on the ADF structure that we have developed for holding ALARA output data, then I think this PR should focus on methods that modify the ADF and then write it to SQL. Everything else is for a specific use case that can come in a future PR that focuses on how to process a variety of data that you have already generated using these new capabilities.

gonuke · 2026-02-04T13:01:11Z

+        lib = aop.DataLibrary()
+        adf = lib.make_entries(run_dicts[run_dict])
+        adf_data.append(adf)
+        adf = pd.concat(adf_data)


Not sure why you need this?

Or maybe it's indented too much and should be outside the loop?

gonuke · 2026-02-04T13:03:29Z

+
+def modify_adf(adf, norm_flux_arr, t_irr_arr, inputs):
+    #Remove some columns:
+    adf.drop(columns=['time', 'time_unit', 'variable', 'var_unit', 'block', 'block_num'], inplace=True)    


Before you do this, will it be safer to first filter to get only the number density info. I guess you are running these to only generate number density, but just in case...

Maybe the same for time?

anu1217 · 2026-04-14T17:38:12Z

What if we limit modify_adf() to include only the correct columns (i.e. drop and rename some, maybe leave flux spectrum shape in), but get rid of the parts that use num pulses, duty cycles, and irradiation time in this PR? The write_to_adf() and write_to_sqlite() methods can stay with some modifications. I think we've deviated somewhat from our original approach since we started focusing on schedule_transforms.py to handle the irradiation time approximations, so I just want to make sure we're on the same page regarding what the scope of this PR is.

gonuke · 2026-04-15T10:46:41Z

I think this should focus on #35 and less on #33. I've left some comments on each of those issues, but for now I think they can largely be decoupled.

gonuke · 2026-04-15T10:57:35Z

To repeat the comments from #35:
I think this PR should assume that a sufficiently complete ADF already exists and not include any functions (write_to_adf()) to generate such a thing. It should focus on:

what columns should exist in your final SQL table
what columns need to be added to/dropped from the ADF
how to write the modified ADF as SQL

Notably, this particular approach for generating the ADF makes sense for the current "campaign" of problems, but may not always make sense as we consider different variations on how we set up problems.

Co-authored-by: Paul Wilson <paul.wilson@wisc.edu>

gonuke

I think it's important to separate the method for knowing the values of t_irr_flat from the modifying of the ADF. More generally, parsing a run_label may become the least common way to know this information.

gonuke · 2026-04-18T15:10:29Z

It looks like you only need this file to get the number of groups. Could that just be an input/command-line parameter? That is, in stead of providing a filename for the group bounds, you can just provide an integer number of groups.

gonuke · 2026-04-18T15:11:13Z

  - 64  
-flux_file : /filespace/a/asrajendra/research/activationDB/ref_flux_files/iter_dt_flux_2.0986E14
+flux_file : ../calc_dwell_dir/ref_flux_files/iter_dt_flux_2.0986E14
+vit_j_file : ../data/vit-j-175-bins.txt


See previous comment - since this file only used to determine the number of groups - maybe this entry is just the number of groups

Suggested change

vit_j_file : ../data/vit-j-175-bins.txt

n_groups : 175

gonuke · 2026-04-18T15:16:46Z

+        inputs = yaml.safe_load(yaml_file)
+    return inputs  
+
+def modify_adf(adf, norm_flux_arr, t_irr_arr, inputs):


Adding docstrings early can help the reviewer know your intent for the data structures.

gonuke · 2026-04-18T15:19:01Z

+    num_pulses = inputs['num_pulses']
+    duty_cycles = inputs['duty_cycles']
+
+    # Extract the number of pulses and duty cycles from the run label
+    pulse_num_dc = adf['run_lbl'].str.extract(r'_(\d+)p_(\d+)_').astype(int)
+
+    # Map num_pulses and duty_cycles to an index
+    pulse_idx = pulse_num_dc[0].map({pulse_num: i for i, pulse_num in enumerate(num_pulses)})
+    duty_cycle_idx  = (pulse_num_dc[1]/100).map({duty_cycle: i for i, duty_cycle in enumerate(duty_cycles)})
+
+    # Add flattened irradiation time:
+    adf['t_irr_flat'] = t_irr_arr.T[pulse_idx.to_numpy(), duty_cycle_idx.to_numpy()]
+    adf['t_irr_flat'] = aop.convert_times(adf['t_irr_flat'], from_unit='y', to_unit='s')


Separation of concerns: All of this seems like it's in the wrong place. All the data for the new columns in the ADF should be computed before you make the call the modify the ADF. Remember, there may be many ways to compute the data for the new columns depending on what the original source of the data is, and you want flexibility to mix and match.

Specifically, I thought we discussed that you shouldn't rely on decoding information from the run_label to get this information, anyway. If you still want to have that as one pathway to determine this data, that should be it's own method with only that purpose.

anu1217 added 4 commits January 16, 2026 14:51

add function to write output files to adf

ea267d8

fix write_to_adf() and make function to write to sqlite

d863350

extend new function

daa624f

add flux spectrum shape and t_irr to db

5cae818

anu1217 added this to the M_01_02: Populate database with the results of M_01_01 milestone Jan 20, 2026

gonuke requested changes Apr 4, 2026

View reviewed changes

anu1217 and others added 8 commits April 17, 2026 11:23

Apply suggestions from code review

242c707

Co-authored-by: Paul Wilson <paul.wilson@wisc.edu>

switch to .read() instead of .readlines()

48fd7e4

Merge branch 'make_adf' of github.com:anu1217/activationDB into make_adf

77c53af

change variable names

cf8dd3a

remove openmc dependency

972d346

store vit j data externally

1099775

move empty flux file check, pass flux entries instead of flux_str

dc11956

remove write_to_adf()

56d610d

gonuke requested changes Apr 18, 2026

View reviewed changes

Conversation

anu1217 commented Jan 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

gonuke left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

anu1217 commented Apr 14, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

gonuke commented Apr 15, 2026

Uh oh!

gonuke commented Apr 15, 2026

Uh oh!

gonuke left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

anu1217 commented Jan 20, 2026 •

edited

Loading

anu1217 commented Apr 14, 2026 •

edited

Loading