Classify machine shop Tier-2 selector outlier

This commit is contained in:
Jan Petykiewicz 2026-04-19 15:54:32 -07:00
commit 7894c52aea
5 changed files with 6714 additions and 278 deletions

View file

@ -96,6 +96,7 @@ pub struct BuildingTypeRecoveredTableSummary {
pub nonzero_bty_header_name_0x40_summaries: Vec<BuildingTypeBtyHeaderNameSummary>,
pub nonzero_bty_header_name_0x5e_summaries: Vec<BuildingTypeBtyHeaderNameSummary>,
pub nonzero_bty_header_name_0x7c_summaries: Vec<BuildingTypeBtyHeaderNameSummary>,
pub nonzero_bty_header_alias_selector_summaries: Vec<BuildingTypeBtyHeaderAliasSelectorSummary>,
pub bty_header_name_0x5e_dword_summaries: Vec<BuildingTypeBtyHeaderNameDwordSummary>,
pub bty_name_0x5e_bca_selector_summaries: Vec<BuildingTypeBtyNameBcaSelectorSummary>,
}
@ -126,6 +127,21 @@ pub struct BuildingTypeBtyHeaderNameDwordSummary {
pub sample_file_names: Vec<String>,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub struct BuildingTypeBtyHeaderAliasSelectorSummary {
pub name_0x40: String,
pub name_0x5e: String,
pub name_0x7c: String,
pub dword_0xbb: u32,
pub dword_0xbb_hex: String,
pub byte_0xb8_hex: String,
pub byte_0xb9_hex: String,
pub byte_0xba_hex: String,
pub byte_0xbb_hex: String,
pub file_count: usize,
pub sample_file_names: Vec<String>,
}
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
pub struct BuildingTypeBtyNameBcaSelectorSummary {
pub header_offset_hex: String,
@ -507,6 +523,8 @@ fn summarize_recovered_table_families(
summarize_nonzero_bty_header_name_lane(files, 0x5e, |probe| &probe.name_0x5e);
let nonzero_bty_header_name_0x7c_summaries =
summarize_nonzero_bty_header_name_lane(files, 0x7c, |probe| &probe.name_0x7c);
let nonzero_bty_header_alias_selector_summaries =
summarize_nonzero_bty_header_alias_selector_patterns(entries, files);
let bty_header_name_0x5e_dword_summaries =
summarize_bty_header_name_lane_by_dword(files, 0x5e, |probe| &probe.name_0x5e);
let bty_name_0x5e_bca_selector_summaries =
@ -528,6 +546,7 @@ fn summarize_recovered_table_families(
nonzero_bty_header_name_0x40_summaries,
nonzero_bty_header_name_0x5e_summaries,
nonzero_bty_header_name_0x7c_summaries,
nonzero_bty_header_alias_selector_summaries,
bty_header_name_0x5e_dword_summaries,
bty_name_0x5e_bca_selector_summaries,
}
@ -700,6 +719,112 @@ fn summarize_bty_name_0x5e_bca_selector_patterns(
summaries
}
fn summarize_nonzero_bty_header_alias_selector_patterns(
entries: &[BuildingTypeSourceEntry],
files: &[BuildingTypeSourceFile],
) -> Vec<BuildingTypeBtyHeaderAliasSelectorSummary> {
let file_by_name = files
.iter()
.map(|file| (file.file_name.as_str(), file))
.collect::<BTreeMap<_, _>>();
let mut groups = BTreeMap::<
(String, String, String, u32, String, String, String, String),
Vec<String>,
>::new();
for entry in entries {
let bty_file = entry
.file_names
.iter()
.filter_map(|name| file_by_name.get(name.as_str()))
.find(|file| matches!(file.source_kind, BuildingTypeSourceKind::Bty));
let bca_file = entry
.file_names
.iter()
.filter_map(|name| file_by_name.get(name.as_str()))
.find(|file| matches!(file.source_kind, BuildingTypeSourceKind::Bca));
let (Some(bty_file), Some(bca_file)) = (bty_file, bca_file) else {
continue;
};
let (Some(bty_probe), Some(bca_probe)) =
(&bty_file.bty_header_probe, &bca_file.bca_selector_probe)
else {
continue;
};
if bty_probe.dword_0xbb == 0 {
continue;
}
let name_0x40 = bty_probe.name_0x40.trim();
let name_0x5e = bty_probe.name_0x5e.trim();
let name_0x7c = bty_probe.name_0x7c.trim();
if name_0x40.is_empty() || name_0x5e.is_empty() || name_0x7c.is_empty() {
continue;
}
groups
.entry((
name_0x40.to_string(),
name_0x5e.to_string(),
name_0x7c.to_string(),
bty_probe.dword_0xbb,
bca_probe.byte_0xb8_hex.clone(),
bca_probe.byte_0xb9_hex.clone(),
bca_probe.byte_0xba_hex.clone(),
bca_probe.byte_0xbb_hex.clone(),
))
.or_default()
.push(bty_file.file_name.clone());
}
let mut summaries = groups
.into_iter()
.map(
|(
(
name_0x40,
name_0x5e,
name_0x7c,
dword_0xbb,
byte_0xb8_hex,
byte_0xb9_hex,
byte_0xba_hex,
byte_0xbb_hex,
),
mut file_names,
)| {
file_names.sort();
file_names.dedup();
BuildingTypeBtyHeaderAliasSelectorSummary {
name_0x40,
name_0x5e,
name_0x7c,
dword_0xbb,
dword_0xbb_hex: format!("0x{dword_0xbb:08x}"),
byte_0xb8_hex,
byte_0xb9_hex,
byte_0xba_hex,
byte_0xbb_hex,
file_count: file_names.len(),
sample_file_names: file_names.into_iter().take(24).collect(),
}
},
)
.collect::<Vec<_>>();
summaries.sort_by(|left, right| {
right
.file_count
.cmp(&left.file_count)
.then_with(|| left.dword_0xbb.cmp(&right.dword_0xbb))
.then_with(|| left.name_0x40.cmp(&right.name_0x40))
.then_with(|| left.name_0x5e.cmp(&right.name_0x5e))
.then_with(|| left.name_0x7c.cmp(&right.name_0x7c))
.then_with(|| left.byte_0xb8_hex.cmp(&right.byte_0xb8_hex))
.then_with(|| left.byte_0xb9_hex.cmp(&right.byte_0xb9_hex))
.then_with(|| left.byte_0xba_hex.cmp(&right.byte_0xba_hex))
.then_with(|| left.byte_0xbb_hex.cmp(&right.byte_0xbb_hex))
});
summaries
}
#[derive(Debug, Clone, PartialEq, Eq, Deserialize)]
struct BuildingBindingArtifact {
bindings: Vec<BuildingBindingRow>,
@ -908,6 +1033,22 @@ mod tests {
sample_file_names: vec!["Port.bty".to_string()],
}]
);
assert_eq!(
summary.nonzero_bty_header_alias_selector_summaries,
vec![BuildingTypeBtyHeaderAliasSelectorSummary {
name_0x40: "Port".to_string(),
name_0x5e: "TextileMill".to_string(),
name_0x7c: "Port".to_string(),
dword_0xbb: 0x01f4,
dword_0xbb_hex: "0x000001f4".to_string(),
byte_0xb8_hex: "0x00".to_string(),
byte_0xb9_hex: "0x00".to_string(),
byte_0xba_hex: "0x00".to_string(),
byte_0xbb_hex: "0x00".to_string(),
file_count: 1,
sample_file_names: vec!["Port.bty".to_string()],
}]
);
assert_eq!(
summary.bty_name_0x5e_bca_selector_summaries,
vec![BuildingTypeBtyNameBcaSelectorSummary {