Added support for float input in scientific notation and removed float rounding.

[sreeramsanjay]

Motivation

There is no support for floating-point inputs with scientific notation in the current implementation of miniWDL. Float values are rounded to 6 decimal points only; for example, floats smaller than 0.000001 are rounded to 0.0. Adding support for float input in scientific notation and removing float rounding using repr method to keep the input float value as is.
NOTE: floats smaller than 0.0001 will be return in scientific notations.

Approach

...

Checklist

• Add appropriate test(s) to the automatic suite
• Use make pretty to reformat the code with black
• Use make check to statically check the code using Pyre and Pylint
• Send PR from a dedicated branch without unrelated edits
• Ensure compatibility with this project's MIT license

[mlin]

@sreeramsanjay Sorry for the very slow uptake on this!

1. The serialization of a float to a string with six decimal places is required by the WDL spec (at least in a command placeholder). Are there other contexts where the serialization could be done differently consistent with the spec?
2. Can you supply a test case where reading a float with an exponent isn't working? I think it should generally, but specific bugs/oversights are possible certainly.

[sreeramsanjay]

@mlin A float with an exponent is throwing an error with the JSON input file. The same is working when passing it as an argument but not with the JSON file(-i).
$ miniwdl run float-test.wdl -i float-input.json
check JSON input; couldn't construct Float? from "1e-30" (in floatToPrint)
refer: https://github.com/sreeramsanjay/wdl/tree/main#

[mlin]

@sreeramsanjay did you try making your input JSON ```json { "floatToPrint": 1e-30 } ``` (without quotes around the number literal)

…

On Tue, Mar 18, 2025 at 4:18 PM ssram ***@***.***> wrote: @mlin <https://github.com/mlin> A float with an exponent is throwing an error with the JSON input file. The same is working when passing it as an argument but not with the JSON file(-i). $ miniwdl run float-test.wdl -i float-input.json check JSON input; couldn't construct Float? from "1e-30" (in floatToPrint) refer: https://github.com/sreeramsanjay/wdl/tree/main# — Reply to this email directly, view it on GitHub <#711 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACXBRXU4YKICIOW63J7E232VDHXNAVCNFSM6AAAAABMQ5EZ2KVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDOMZVGE2TENZQGA> . You are receiving this because you were mentioned.Message ID: ***@***.***> [image: sreeramsanjay]*sreeramsanjay* left a comment (chanzuckerberg/miniwdl#711) <#711 (comment)> @mlin <https://github.com/mlin> A float with an exponent is throwing an error with the JSON input file. The same is working when passing it as an argument but not with the JSON file(-i). $ miniwdl run float-test.wdl -i float-input.json check JSON input; couldn't construct Float? from "1e-30" (in floatToPrint) refer: https://github.com/sreeramsanjay/wdl/tree/main# — Reply to this email directly, view it on GitHub <#711 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AACXBRXU4YKICIOW63J7E232VDHXNAVCNFSM6AAAAABMQ5EZ2KVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDOMZVGE2TENZQGA> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[sreeramsanjay]

@mlin yes, I tried and it's the same error.

[mlin]

@sreeramsanjay Thanks, I tracked that down to a longstanding quirk in PyYAML (which miniwdl uses to read the input file, JSON in theory being a subset of YAML): yaml/pyyaml#173 .

As a workaround, if you write 1.0e-30 instead of 1e-30 then it's recognized as expected.

Code options we'll consider:

• overriding PyYAML's float-recognizing regexes: https://stackoverflow.com/a/30462009
• switch back to ruamel.yaml, an alternative YAML package that doesn't have this quirk. (which we've used in the past; the better-maintained of the two seems to oscillate every few years)