docs: spec + plan to recompute OD from VTank loot profile (0-10 scale)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
Erik 2026-07-15 19:06:02 +02:00
parent cc403240d6
commit 7a11122a26
2 changed files with 380 additions and 0 deletions

View file

@ -0,0 +1,303 @@
# Weapon OD from Loot Profile — Implementation Plan
> **For agentic workers:** REQUIRED SUB-SKILL: superpowers:subagent-driven-development. Steps use checkbox (`- [ ]`) syntax.
**Goal:** Replace the OD calculation with the user's VTank loot-profile ladder so `od_rating` is the 010 (casters 07) integer they see on ident. Spec: `docs/superpowers/specs/2026-07-15-weapon-od-loot-profile-design.md`.
**Architecture:** Rewrite `go-services/inventory-go/od.go` around an ordered bucket table (below, extracted from `Loot.utl`). `OD = clamp(stat baseline, 1, cap)` for the first matching bucket. Backfill; the frontend renders integer OD.
**Tech Stack:** Go 1.25 (server-only tests in `golang:1.25-bookworm`), React (npm build).
**Working dir for git:** `C:/Users/erikn/source/repos/dereth-workspace/MosswartOverlord`
**Go test command:**
```bash
cd "C:/Users/erikn/source/repos/dereth-workspace/MosswartOverlord" && \
tar czf - go-services | ssh erik@overlord.snakedesert.se \
"rm -rf /tmp/tdd-verify && mkdir -p /tmp/tdd-verify && tar xzf - -C /tmp/tdd-verify" && \
ssh erik@overlord.snakedesert.se \
"docker run --rm -v /tmp/tdd-verify/go-services/inventory-go:/src -w /src golang:1.25-bookworm \
sh -c 'go mod tidy >/dev/null 2>&1; go test ./... -run <PATTERN> -v 2>&1'"
```
**Item keys:** object_class from `raw["ObjectClass"]`; skill `IntValues["218103840"]` (melee/missile) / `IntValues["159"]` (casters); mastery `IntValues["353"]`; MaxDamage `IntValues["218103842"]`; missile dmg `IntValues["218103839"]`; elemVsMon `DoubleValues["152"]`; name `raw["Name"]`. Helpers `bag/ivI/dvF/rawI/rawS` exist in process.go.
---
### Task 1: Rewrite od.go with the loot-profile table (TDD)
**Files:** rewrite `go-services/inventory-go/od.go`; rewrite `go-services/inventory-go/od_test.go`.
- [ ] **Step 1: Write tests first** (`od_test.go`, replacing the old content):
```go
package main
import "testing"
func TestOD_MeleeTetsubo(t *testing.T) {
// 2H Cleaving baseline 45, name "Tetsubo", maxdmg 85 → clamp(85-45,1,10)=10.
raw := map[string]any{"ObjectClass": 1.0, "Name": "Tetsubo",
"IntValues": map[string]any{"218103840": 41.0, "353": 11.0, "218103842": 85.0}}
if od, ok := computeOD(raw); !ok || od != 10 {
t.Fatalf("Tetsubo OD=%v ok=%v want 10", od, ok)
}
}
func TestOD_MeleeMidTier(t *testing.T) {
// Heavy Axe baseline 74, name "War Axe", maxdmg 79 → clamp(79-74,1,10)=5.
raw := map[string]any{"ObjectClass": 1.0, "Name": "Bronze War Axe",
"IntValues": map[string]any{"218103840": 44.0, "353": 3.0, "218103842": 79.0}}
if od, ok := computeOD(raw); !ok || od != 5 {
t.Fatalf("War Axe OD=%v ok=%v want 5", od, ok)
}
}
func TestOD_MeleeBelowBaselineNull(t *testing.T) {
// Heavy Axe baseline 74, maxdmg 74 → stat-baseline=0 (<1) null.
raw := map[string]any{"ObjectClass": 1.0, "Name": "Bronze War Axe",
"IntValues": map[string]any{"218103840": 44.0, "353": 3.0, "218103842": 74.0}}
if _, ok := computeOD(raw); ok {
t.Fatal("maxdmg == baseline must be null")
}
}
func TestOD_MeleeNameGateNull(t *testing.T) {
// Heavy Axe bucket requires name in {Silifi,Lugian,War Axe,Battle}. A skill-44
// mastery-3 item NOT matching any name and matching no other bucket → null.
raw := map[string]any{"ObjectClass": 1.0, "Name": "Bronze Frobnicator",
"IntValues": map[string]any{"218103840": 44.0, "353": 3.0, "218103842": 84.0}}
if _, ok := computeOD(raw); ok {
t.Fatal("name not matching the Heavy Axe filter must be null")
}
}
func TestOD_Crossbow(t *testing.T) {
// Xbow baseline 77, k839=89 → clamp(89-77,1,10)=10.
raw := map[string]any{"ObjectClass": 9.0, "Name": "Fire Compound Crossbow",
"IntValues": map[string]any{"218103840": 47.0, "353": 9.0, "218103839": 89.0}}
if od, ok := computeOD(raw); !ok || od != 10 {
t.Fatalf("Xbow OD=%v ok=%v want 10", od, ok)
}
}
func TestOD_WarCaster(t *testing.T) {
// War caster: elemVsMon 1.40 → clamp(round((1.40-1.18)*100),1,7)=clamp(22,1,7)=7.
raw := map[string]any{"ObjectClass": 31.0, "Name": "Diamond Frost Baton",
"IntValues": map[string]any{"159": 34.0},
"DoubleValues": map[string]any{"152": 1.40}}
if od, ok := computeOD(raw); !ok || od != 7 {
t.Fatalf("War caster OD=%v ok=%v want 7", od, ok)
}
}
func TestOD_VoidCasterTier(t *testing.T) {
// Void caster: name must contain "Nether"; elemVsMon 1.22 → round((1.22-1.18)*100)=4.
raw := map[string]any{"ObjectClass": 31.0, "Name": "Silver Nether Staff",
"IntValues": map[string]any{"159": 43.0},
"DoubleValues": map[string]any{"152": 1.22}}
if od, ok := computeOD(raw); !ok || od != 4 {
t.Fatalf("Void caster OD=%v ok=%v want 4", od, ok)
}
}
func TestOD_BowDeferredNull(t *testing.T) {
// Bows (mastery 8) are deferred → always null in this pass.
raw := map[string]any{"ObjectClass": 9.0, "Name": "Corsair's Arc",
"IntValues": map[string]any{"218103840": 47.0, "353": 8.0, "218103839": 70.0}}
if _, ok := computeOD(raw); ok {
t.Fatal("bows are deferred → null")
}
}
```
- [ ] **Step 2: Run → RED** (`-run TestOD`): fails (old computeOD signature returns float; these expect int). Build/logic failure expected.
- [ ] **Step 3: Rewrite `od.go`** (delete ALL old content — table, spell dicts, variance math — and replace):
```go
package main
import "strings"
// Weapon OD = the user's VirindiTank Loot.utl classification (010, casters
// 07), printed by Mag-Tools on ident. OD = clamp(stat baseline, 1, cap) for
// the first matching bucket, else null. Buckets/baselines transcribed from
// Loot.utl (see design doc 2026-07-15-weapon-od-loot-profile).
type odBucket struct {
objectClass int // 1 melee, 9 missile, 31 caster
skill int // 218103840 for melee/missile, 159 for casters; 0 = any
mastery int // IntValues 353; -1 = not checked
nameAlts []string // OR of case-sensitive substrings; empty = no name gate
statKey string // IntValues key (melee/missile) or ""
dblStatKey string // DoubleValues key (casters) or ""
baseline float64
cap int
}
// Ordered: name-gated buckets are effectively specific; first match wins.
var odBuckets = []odBucket{
// --- Melee (object_class 1, stat MaxDamage 218103842, cap 10) ---
{1, 44, 3, []string{"Silifi", "Lugian", "War Axe", "Battle"}, "218103842", "", 74, 10}, // (H) Axe
{1, 45, 3, []string{"Dolabra", "Ono", "Hand", "War Hammer"}, "218103842", "", 61, 10}, // (L) Axe
{1, 46, 3, []string{"Hatchet", "Shou-ono", "Tungi", "Hammer"}, "218103842", "", 61, 10}, // (F) Axe
{1, 44, 6, []string{"Stiletto", "Jambiya"}, "218103842", "", 38, 10}, // (H) Dagger MS
{1, 44, 6, nil, "218103842", "", 71, 10}, // (H) Dagger
{1, 45, 6, nil, "218103842", "", 58, 10}, // (L) Dagger (MS shares names=nil; see note)
{1, 46, 6, []string{"Knife", "Lancet"}, "218103842", "", 28, 10}, // (F) Dagger MS
{1, 46, 6, nil, "218103842", "", 58, 10}, // (F) Dagger
{1, 44, 4, []string{"Mazule", "Mace", "Morning"}, "218103842", "", 69, 10}, // (H) Mace
{1, 45, 4, []string{"Club", "Kasrullah"}, "218103842", "", 56, 10}, // (L) Mace
{1, 46, 4, []string{"Board", "Tofun", "Dabus"}, "218103842", "", 56, 10}, // (F) Mace
{1, 44, 5, []string{"Glaive", "Trident", "Partizan"}, "218103842", "", 72, 10}, // (H) Spear
{1, 45, 5, []string{"Spear", "Yari"}, "218103842", "", 60, 10}, // (L) Spear
{1, 44, 7, []string{"Nabut", "Stick"}, "218103842", "", 69, 10}, // (H) Staff
{1, 45, 7, nil, "218103842", "", 57, 10}, // (L) Staff
{1, 46, 7, []string{"Jo", "Bastone"}, "218103842", "", 57, 10}, // (F) Staff
{1, 44, 2, []string{"Takuba", "Flamberge", "Ken", "Long", "Tachi"}, "218103842", "", 71, 10}, // (H) Sword
{1, 45, 2, []string{"Dericost", "Broad", "Shamshir", "Kaskara", "Spada"}, "218103842", "", 58, 10}, // (L) Sword
{1, 46, 2, []string{"Scimitar", "Yaoji", "Short", "Sabra", "Simi"}, "218103842", "", 58, 10}, // (F) Sword
{1, 44, 1, []string{"Nekode", "Cestus"}, "218103842", "", 59, 10}, // (H) UA
{1, 45, 1, []string{"Katar", "Knuckles"}, "218103842", "", 48, 10}, // (L) UA
{1, 46, 1, []string{"Claw", "Wraps"}, "218103842", "", 48, 10}, // (F) UA
{1, 41, -1, []string{"Nodachi", "Shashqa", "Spadone", "Greataxe", "Quadrelle", "Khanda-handled", "Tetsubo", "Star"}, "218103842", "", 45, 10}, // (2H) Cleaving
{1, 41, -1, []string{"Assagai", "Pike", "Magari", "Corsesca"}, "218103842", "", 48, 10}, // (2H) Spear
// --- Crossbow (object_class 9, mastery 9, stat 218103839, cap 10) ---
{9, 47, 9, nil, "218103839", "", 77, 10},
// --- Casters (object_class 31, stat elemVsMon 152, cap 7) ---
{31, 34, -1, []string{"Baton", "Sceptre", "Staff"}, "", "152", 1.18, 7}, // War
{31, 43, -1, []string{"Nether"}, "", "152", 1.18, 7}, // Void
// Bows (mastery 8) and Thrown (mastery 10) are intentionally ABSENT → null.
}
func nameMatches(name string, alts []string) bool {
if len(alts) == 0 {
return true
}
for _, a := range alts {
if strings.Contains(name, a) {
return true
}
}
return false
}
// computeOD returns the loot-profile OD (integer, as float64 for the DB column)
// and ok=false when no bucket matches or the stat is at/below baseline.
func computeOD(raw map[string]any) (float64, bool) {
oc := rawI(raw, "ObjectClass", 0)
iv := bag(raw, "IntValues")
dv := bag(raw, "DoubleValues")
name := rawS(raw, "Name")
skillMelee := ivI(iv, "218103840", 0)
skillCaster := ivI(iv, "159", 0)
mastery := ivI(iv, "353", -999)
for _, b := range odBuckets {
if b.objectClass != oc {
continue
}
sk := skillMelee
if b.objectClass == 31 {
sk = skillCaster
}
if b.skill != 0 && b.skill != sk {
continue
}
if b.mastery != -1 && b.mastery != mastery {
continue
}
if !nameMatches(name, b.nameAlts) {
continue
}
var stat float64
if b.dblStatKey != "" {
stat = dvF(dv, b.dblStatKey, 0)
} else {
stat = float64(ivI(iv, b.statKey, 0))
}
od := stat - b.baseline
if b.dblStatKey != "" {
od = od * 100 // caster elemVsMon: percentage points
}
odInt := int(od + 0.5) // round
if odInt < 1 {
return 0, false
}
if odInt > b.cap {
odInt = b.cap
}
return float64(odInt), true
}
return 0, false
}
```
Note on the (L) Dagger MS bucket: the profile has an `(L) Dagger MS` (baseline 28) with no name filter, which would collide with `(L) Dagger` (baseline 58, also no name). Two nameless same-skill/mastery buckets can't be distinguished by our fields; the plan keeps only `(L) Dagger` (baseline 58) — the more conservative (higher-baseline, fewer false OD10s). Document this as a known minor gap in the commit message.
- [ ] **Step 4: Run → GREEN** (`-run TestOD`) all pass; then `-run .` whole package + `go vet ./...` clean.
- [ ] **Step 5: Commit**
```
feat(inventory-go): recompute OD from VTank loot-profile ladder (0-10 / casters 0-7)
Replaces the UtilityBelt over-retail float with the user's actual in-game OD:
per-weapon-type MaxDamage/elemVsMon thresholds from Loot.utl. Bows/thrown
deferred (null) pending a computed-stat calibration. (L) Dagger MS bucket
folded into (L) Dagger (indistinguishable by captured fields).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
```
---
### Task 2: Frontend integer rendering
**Files:** `frontend/src/components/inventory/ResultsTable.tsx`, `DetailPanel.tsx`.
- [ ] **Step 1:** In `ResultsTable.tsx` `cell()` `case 'od_rating'`, render integer:
```ts
case 'od_rating': {
const v = item.od_rating;
return v == null ? '—' : String(v);
}
```
- [ ] **Step 2:** In `DetailPanel.tsx`, change the OD row to plain integer:
```tsx
{item.od_rating != null && <Row k="Weapon OD" v={String(item.od_rating)} />}
```
- [ ] **Step 3:** Build `cd frontend && npm run build`; delete `static/_build`. Commit:
```
feat(frontend): render weapon OD as integer (0-10 loot-profile scale)
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
```
---
### Task 3: Deploy backend + backfill + verify
- [ ] **Step 1:** Sync + test-gated build + recreate inventory-go (cutover override) — standard flow from CLAUDE.md.
- [ ] **Step 2:** Backfill: `ssh erik@overlord.snakedesert.se "curl -s -X POST http://127.0.0.1:8772/admin/backfill-od"`. Expect `{scanned, updated}`.
- [ ] **Step 3:** Verify distribution is 010 / casters ≤7 and spot-check known items:
```bash
ssh erik@overlord.snakedesert.se "docker exec inventory-db psql -U inventory_user -d inventory_db -tc \"SELECT min(od_rating),max(od_rating),count(*) FILTER (WHERE od_rating IS NOT NULL) FROM items WHERE object_class IN (1,9,31);\""
ssh erik@overlord.snakedesert.se "curl -s 'http://127.0.0.1:8770/inv/search/items?include_all_characters=true&weapon_only=true&min_od=1&sort_by=od&sort_dir=desc&limit=10' | python3 -c \"import json,sys;d=json.load(sys.stdin);[print(i.get('od_rating'),'|',i['name']) for i in d['items']]\""
```
Expect max = 10, casters max 7, Tetsubo = 10.
---
### Task 4: Deploy frontend + user verify
- [ ] `bash deploy-frontend.sh`, commit `static/`+`frontend/`, push, server `git pull` (handle the tar-sync working-tree blocker as before: verify blockers match origin, checkout/rm, re-pull).
- [ ] **User check:** ID a few weapons in-game and confirm the site's OD column matches the grey OD text (melee + crossbows + casters). Bows/thrown show `—` (deferred) — confirm that's acceptable and, if wanted, provide one bow OD reading to finish them.
---
## Self-review
- Melee 25 buckets + xbow + 2 casters implemented; bows/thrown null (documented).
- (L) Dagger MS folded into (L) Dagger — noted.
- Old variance/retail od.go fully deleted.
- od_rating stays the same column/param/sort; only values + rendering change.

View file

@ -0,0 +1,77 @@
# Weapon OD — recompute from the VTank loot profile (010 scale)
**Date:** 2026-07-15
**Status:** Approved (user directed the rebuild)
## Problem
The shipped OD (`od.go`) ports UtilityBelt's unbounded "over-retail damage"
float (16 … +43). The user's actual in-game OD is a **010** (casters 07)
integer they see on ident. Investigation (this session) proved:
- Mag-Tools has **no** OD calculation. Its `ItemInfoPrinter` prints the item
info **plus the matching VirindiTank loot-rule name** (`GetLootRuleInfoFromItemInfo`).
- The user's `VirindiTank\Loot.utl` defines a ladder of rules named
`(H) Axe (OD +10)`, `(OD +9)`, … per weapon bucket. Mag-Tools prints whichever
matched — **that** is the OD the user sees.
The thresholds are the same across all the user's profiles (user-confirmed), so
one table drives every weapon and caster.
## The rule (decoded from Loot.utl)
Each bucket is `OD = clamp(stat baseline, 1, cap)`, where the bucket is matched
by **weapon skill + mastery + (optional) weapon-name filter**, and `stat` is a
raw captured value. Ladders step by 1 per tier.
- **Melee** (object_class 1): `stat = MaxDamage` (int key `218103842`), cap 10.
27 buckets (skill `218103840`, mastery `353`, name filter, baseline) — see the
table in the plan. Example: 2H Cleaving baseline 45 → Tetsubo (maxdmg 85) = OD
**+10** (matches in-game); Heavy Axe baseline 74.
- **Crossbow** (object_class 9, mastery `353`=9): `stat = int key 218103839`
("effective missile damage"), baseline **77**, cap 10 (OD+10=87 … OD+1=78,
verified on Fire Compound Crossbow k839=89 → +10).
- **Casters** (object_class 31): `stat = ElementalDamageVersusMonsters` (double
key `152`), baseline **1.18**, ×100, cap **7**. War bucket: skill `159`=34,
name `Baton|Sceptre|Staff`; Void bucket: skill 43, name `Nether`. OD+7 at
elemVsMon 1.25, OD+1 at 1.19. Frost Baton (1.40) → +7.
- **No match / stat below baseline+1 → NULL** (item shows no OD, exactly like
in-game when no OD rule matches).
## Deferred: Bows & Thrown (mastery 8 / 10)
Their OD rules threshold a **computed decimal** damage stat (bow OD+10 = 78.6,
thrown = 85.3) on a key not reliably identifiable from the profile alone (bow
`218103839` values 4570 don't reach 78.6; the rule uses a DoubleValKeyGE the
`.utl` encodes indirectly). To avoid showing a *wrong* number (the very problem
we're fixing), **bows and thrown get NULL OD** in this pass. Reopened once the
user supplies one in-game bow OD reading to calibrate the stat.
## Design
- **`od.go` rewritten** around an ordered bucket table (extracted from
`Loot.utl`, committed as data): `{objectClass, skill, mastery, nameAlts[],
statKey, baseline, cap}`. `computeOD` finds the first bucket whose
objectClass+skill+mastery+name match, then `OD = clamp(stat baseline, 1,
cap)`; returns null if none match or stat ≤ baseline. Name match = any
alternative is a substring of the item name (VTank substring semantics).
- **Type**: `od_rating` stays `DOUBLE PRECISION NULL` but now holds an integer
110 (or 17). Frontend already renders it; drop the `+`/decimals for the new
integer (show plain `7`, `10`, `—`).
- Search `min_od`/`max_od`/sort unchanged. Backfill via existing
`POST /admin/backfill-od`.
- The best-values table, spell-effect dicts, and variance math in the old
`od.go` are **deleted** (wrong metric).
## Testing
Go golden tests from real items: Tetsubo → 10, Frost Baton → 7, Fire Compound
Crossbow → 10, a mid melee → its expected tier, a below-baseline weapon → null,
a bow → null (deferred). Then live backfill + spot-checks; the column must read
010/07 and the user confirms against ident.
## Out of scope
- Bows/thrown OD (deferred, null).
- The loot-rule name filters are reproduced as substring lists; exotic renamed
items may differ from a live VTank match — acceptable.