Struct ObservationLog 

pub struct ObservationLog {
    pub observations: Vec<Observation>,
}

The append-only feedback log for one taste profile.

Fields

observations: Vec<Observation>

The observations, in arrival order.

Read row by row, and a row that will not parse is skipped rather than failing the log. That is a one-line attribute standing in for a very large failure: this Vec is nested inside Profile, which is nested inside SessionState, so a single unreadable cell anywhere in the log used to abort the whole deserialize — and the caller’s only recovery is to start a new session. Losing one vote is a rounding error; losing the bank, the lineage, the pins and every vote because of it is the exact opposite of the promise the raw-φ log was built on. Observed, not imagined: one null in one φ vector silently replaced a forty-patch, gen-39 session with an empty one.

It is deliberately not quiet about the cost — Self::load and the session layer both re-serialize what they read, so a skipped row is gone from disk on the next save. That is the right trade only because a row that cannot be parsed cannot be interpreted either; there is nothing in it to keep.

Implementations

impl ObservationLog

pub fn new() -> Self

An empty log.

pub fn push(&mut self, obs: Observation)

Append an observation.

pub fn len(&self) -> usize

Number of observations.

pub fn is_empty(&self) -> bool

True when the log holds no observations.

pub fn n_sessions(&self) -> usize

Number of distinct sessions referenced (max session index + 1).

pub fn n_with(&self, provenance: Provenance) -> usize

How many observations were collected each way. A count, not a weight: nothing downstream of the likelihood reads it, and the panel shows it so “the model learned this from a heard comparison” and “…from a checkbox” are distinguishable claims on screen as well as in the log.

pub fn raw_rows(&self, names: &[String]) -> Vec<Vec<f64>>

Every raw φ in the log, for fitting a standardizer. Only observations already in the current schema whose names match names contribute — mixing units into a standardizer is how you get a silently wrong model.

pub fn save(&self, path: &Path) -> Result<()>

Serialize to a JSON file.

pub fn load(path: &Path) -> Result<Self>

Load from a JSON file.

Trait Implementations

impl Clone for ObservationLog

fn clone(&self) -> ObservationLog

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for ObservationLog

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for ObservationLog

fn default() -> ObservationLog

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for ObservationLog

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for ObservationLog

fn eq(&self, other: &ObservationLog) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for ObservationLog

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for ObservationLog