Repository navigation
Simplify float stat conversion using Pydantic coercion #340
Description
Summary
PR #241 adds a custom BeforeValidator / OptionalFloat type to convert MLB Stats API string values such as ".287" into Python floats.
Pydantic already performs this coercion for fields typed as float or Optional[float] (float | None), so the custom conversion logic is unnecessary for normal numeric strings.
Related PR: #241
Proposed change
Replace uses of the custom OptionalFloat type with standard Pydantic annotations, for example:
avg: Optional[float] = None
ops: Optional[float] = None
slg: Optional[float] = NonePydantic will convert values such as:
".287" -> 0.287
"1.024" -> 1.024without a custom validator.
MLB sentinel values
Before removing the validator entirely, verify whether the MLB Stats API can return non-numeric sentinel values for these fields, such as:
.---
-
""
If those values occur, normalization should handle only those MLB-specific sentinel values and allow Pydantic to perform the normal string-to-float coercion.
We should avoid silently converting every invalid value to None, since that can hide unexpected API data or model errors.
Acceptance criteria
- Verify representative MLB responses for numeric float fields and identify any sentinel/non-numeric values.
- Replace
OptionalFloatwithOptional[float]/float | Nonewhere standard Pydantic coercion is sufficient. - Remove the generic
float_or_none/BeforeValidatorconversion if it is no longer needed. - If MLB sentinel values require normalization, handle only the known sentinel values explicitly.
- Add or update tests covering numeric strings,
None, and any known MLB sentinel values. - Ensure the full test suite passes.
Activity
Mattsface commentedon Sep 3, 2026
Fields to update from PR #241
The following string-backed stat fields are candidates to replace with standard
Optional[float]/float | None, allowing Pydantic to handle numeric-string coercion directly.SimpleCatchingSplitavgobpslgopscaught_stealing_percentagestolen_base_percentagestrikeout_walk_ratioSimpleFieldingSplitcaught_stealing_percentagestolen_base_percentagefieldingrange_factor_per_gamerange_factor_per_9_inncatcher_eraAdvancedHittingSplitbabippitches_per_plate_appearancewalks_per_plate_appearancestrikeouts_per_plate_appearancehome_runs_per_plate_appearancewalks_per_strikeoutisoSimpleHittingSplitavgobpslgopscaught_stealing_percentagestolen_base_percentagebabipgroundouts_to_airoutsat_bats_per_home_runSimplePitchingSplitavgobpslgopscaught_stealing_percentagestolen_base_percentageerawhipstrike_percentagegroundouts_to_airoutswin_percentagepitches_per_inningstrikeout_walk_ratiostrikeouts_per_9_innwalks_per_9_innhits_per_9_innruns_scored_per_9home_runs_per_9AdvancedPitchingSplitwinning_percentageruns_scored_per_9babipobpslgopsstrikeouts_per_9base_on_balls_per_9home_runs_per_9hits_per_9strikeouts_to_walksstrikeouts_minus_walks_percentagegidp_percentagebatters_faced_per_gamewhiff_percentagestrike_percentagepitches_per_inningpitches_per_plate_appearancewalks_per_plate_appearancestrikeouts_per_plate_appearancehome_runs_per_plate_appearancewalks_per_strikeoutisoflyball_percentageExpectedStatisticsavgslgwobawobaconThe baseball-notation innings fields are intentionally excluded from this conversion list:
SimpleFieldingSplit.inningsSimplePitchingSplit.innings_pitchedAdvancedPitchingSplit.innings_pitched_per_gameThey should remain outside the scope of this issue because values such as
"6.2"represent 6 2/3 innings rather than the mathematical decimal 6.2.Normalize MLB float sentinel strings before Pydantic coercion
Convert numeric stat fields from str to float via Pydantic coercion