Merge pull request #2760 from NicolasKion/feature/damage-projection-graph

DarkFenX · web-flow · commit 0469e9733562 · 2026-05-21T22:03:14.000+02:00
Add damage projection graph
diff --git a/graphs/data/__init__.py b/graphs/data/__init__.py
@@ -19,6 +19,7 @@
 
 
 from . import fitDamageStats
+from . import fitDamageEnvelope
 from . import fitEwarStats
 from . import fitRemoteReps
 from . import fitShieldRegen
diff --git a/graphs/data/fitDamageEnvelope/__init__.py b/graphs/data/fitDamageEnvelope/__init__.py
@@ -0,0 +1,23 @@
+# =============================================================================
+# Copyright (C) 2010 Diego Duclos
+#
+# This file is part of pyfa.
+#
+# pyfa is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# pyfa is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with pyfa.  If not, see <http://www.gnu.org/licenses/>.
+# =============================================================================
+
+
+from .graph import FitDamageEnvelopeGraph
+
+FitDamageEnvelopeGraph.register()
diff --git a/graphs/data/fitDamageEnvelope/getter.py b/graphs/data/fitDamageEnvelope/getter.py
@@ -0,0 +1,317 @@
+# =============================================================================
+# Copyright (C) 2010 Diego Duclos
+#
+# This file is part of pyfa.
+#
+# pyfa is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# pyfa is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with pyfa.  If not, see <http://www.gnu.org/licenses/>.
+# =============================================================================
+
+
+import eos.config
+from eos.const import FittingHardpoint
+from eos.saveddata.targetProfile import TargetProfile
+from eos.utils.spoolSupport import SpoolOptions, SpoolType
+from graphs.calc import checkLockRange
+from graphs.data.base import SmoothPointGetter
+from graphs.data.fitDamageStats.calc.application import (_calcMissileFactor, _calcTurretChanceToHit, _calcTurretMult,
+                                                         getApplicationPerKey, )
+from service.settings import GraphSettings
+
+
+def _buildResistProfile(tgtResists, tgtFullHp):
+    if not GraphSettings.getInstance().get('ignoreResists'):
+        emRes, thermRes, kinRes, exploRes = tgtResists
+    else:
+        emRes = thermRes = kinRes = exploRes = 0
+    return TargetProfile(emAmount=emRes, thermalAmount=thermRes, kineticAmount=kinRes, explosiveAmount=exploRes,
+        hp=tgtFullHp)
+
+
+def _typedDmgScalar(dmgTyped, applicationMult, profile):
+    """Apply application multiplier and resist profile, return scalar EHP/s."""
+    if applicationMult == 0:
+        return 0
+    scaled = dmgTyped * applicationMult
+    scaled.profile = profile
+    return scaled.total
+
+
+def _turretApplication(snapshot, src, tgt, atkSpeed, atkAngle, distance, tgtSpeed, tgtAngle, tgtSigRadius):
+    cth = _calcTurretChanceToHit(atkSpeed=atkSpeed, atkAngle=atkAngle, atkRadius=src.getRadius(),
+        atkOptimalRange=snapshot['maxRange'] or 0, atkFalloffRange=snapshot['falloff'] or 0,
+        atkTracking=snapshot['tracking'], atkOptimalSigRadius=snapshot['optimalSigRadius'], distance=distance,
+        tgtSpeed=tgtSpeed, tgtAngle=tgtAngle, tgtRadius=tgt.getRadius(), tgtSigRadius=tgtSigRadius)
+    return _calcTurretMult(cth)
+
+
+def _missileApplication(snapshot, distance, tgtSpeed, tgtSigRadius):
+    rangeData = snapshot['missileMaxRangeData']
+    if rangeData is None:
+        return 0
+    lowerRange, higherRange, higherChance = rangeData
+    if distance is None or distance <= lowerRange:
+        distanceFactor = 1
+    elif lowerRange < distance <= higherRange:
+        distanceFactor = higherChance
+    else:
+        distanceFactor = 0
+    if distanceFactor == 0:
+        return 0
+    applicationFactor = _calcMissileFactor(atkEr=snapshot['aoeCloudSize'], atkEv=snapshot['aoeVelocity'],
+        atkDrf=snapshot['aoeDamageReductionFactor'], tgtSpeed=tgtSpeed, tgtSigRadius=tgtSigRadius)
+    return distanceFactor * applicationFactor
+
+
+def _snapshotTurret(mod, dmgTyped, charge):
+    return {'kind': 'turret', 'charge': charge, 'dmg': dmgTyped, 'maxRange': mod.maxRange, 'falloff': mod.falloff,
+        'tracking': mod.getModifiedItemAttr('trackingSpeed'),
+        'optimalSigRadius': mod.getModifiedItemAttr('optimalSigRadius')}
+
+
+def _snapshotMissile(mod, dmgTyped, charge):
+    return {'kind': 'missile', 'charge': charge, 'dmg': dmgTyped, 'missileMaxRangeData': mod.missileMaxRangeData,
+        'aoeCloudSize': mod.getModifiedChargeAttr('aoeCloudSize'),
+        'aoeVelocity': mod.getModifiedChargeAttr('aoeVelocity'),
+        'aoeDamageReductionFactor': mod.getModifiedChargeAttr('aoeDamageReductionFactor'),
+        'isFoF': 'fofMissileLaunching' in (charge.effects if charge else {})}
+
+
+def _isAmmoEnvelopeWeapon(mod):
+    """Turret or standard missile launcher with valid charges."""
+    if mod.hardpoint not in (FittingHardpoint.TURRET, FittingHardpoint.MISSILE):
+        return False
+    # Skip exotic weapon groups handled separately by stock app logic
+    if mod.item.group.name in ('Missile Launcher Bomb', 'Structure Guided Bomb Launcher'):
+        return False
+    if 'ChainLightning' in mod.item.effects:
+        return False
+    if mod.isBreacher:
+        return False
+    return bool(mod.getValidCharges())
+
+
+def _snapshotForCurrentCharge(mod):
+    """Build a snapshot dict for whatever charge is currently loaded on mod."""
+    spoolOptions = SpoolOptions(SpoolType.SPOOL_SCALE, eos.config.settings['globalDefaultSpoolupPercentage'], False)
+    dmgTyped = mod.getDps(spoolOptions=spoolOptions)
+    if mod.hardpoint == FittingHardpoint.TURRET:
+        return _snapshotTurret(mod, dmgTyped, mod.charge)
+    return _snapshotMissile(mod, dmgTyped, mod.charge)
+
+
+def _collectWeaponCandidates(src):
+    """For each ammo-bearing weapon, return list of per-charge snapshots.
+
+    Charge-dependent attributes (optimal/falloff/tracking/missile range/AoE) are
+    only applied to the module's modified attributes by a full fit recalc.
+    Since ammo effects are gun-local in EVE (a crystal in laser-1 does not
+    affect laser-2's attributes), we load up to N different ammos onto N
+    different weapons of the same group, recalc the fit once, and snapshot
+    all N (weapon, ammo) pairs from that single recalc. For a group of size
+    K weapons and M ammos this needs ceil(M / K) recalcs instead of M.
+    Originals are always restored via try/finally even if a calc raises.
+    """
+    fit = src.item
+    weapon_mods = [mod for mod in fit.activeModulesIter() if _isAmmoEnvelopeWeapon(mod)]
+    if not weapon_mods:
+        return []
+
+    # Group by (item ID, state) — within such a group, snapshots can be shared
+    # across mods, and DPS reads need consistent per-mod state.
+    groups = {}
+    for mod in weapon_mods:
+        groups.setdefault((mod.item.ID, mod.state), []).append(mod)
+
+    originals = {id(mod): mod.charge for mod in weapon_mods}
+    snapshots_by_mod = {id(mod): [] for mod in weapon_mods}
+    spoolOptions = SpoolOptions(SpoolType.SPOOL_SCALE, eos.config.settings['globalDefaultSpoolupPercentage'], False)
+
+    try:
+        for group_mods in groups.values():
+            valid_charges = sorted(group_mods[0].getValidCharges(), key=lambda c: c.name)
+            if not valid_charges:
+                continue
+            chunk_size = len(group_mods)
+            for chunk_start in range(0, len(valid_charges), chunk_size):
+                chunk = valid_charges[chunk_start:chunk_start + chunk_size]
+                # Assign one chunk-ammo per group mod (extras stay on their previous charge)
+                for i, charge in enumerate(chunk):
+                    group_mods[i].charge = charge
+                fit.clear()
+                fit.calculateModifiedAttributes()
+                # Snapshot per (assignee mod, charge); copy to all group mods since
+                # within an (item ID, state) group attributes for a given ammo match.
+                for i, charge in enumerate(chunk):
+                    assignee = group_mods[i]
+                    dmgTyped = assignee.getDps(spoolOptions=spoolOptions)
+                    if dmgTyped.total <= 0:
+                        continue
+                    if assignee.hardpoint == FittingHardpoint.TURRET:
+                        snap = _snapshotTurret(assignee, dmgTyped, charge)
+                    else:
+                        snap = _snapshotMissile(assignee, dmgTyped, charge)
+                    for target_mod in group_mods:
+                        snapshots_by_mod[id(target_mod)].append(snap)
+    finally:
+        for mod in weapon_mods:
+            mod.charge = originals[id(mod)]
+        fit.clear()
+        fit.calculateModifiedAttributes()
+
+    weapons = [{'mod': mod, 'candidates': snapshots_by_mod[id(mod)]} for mod in weapon_mods if
+               snapshots_by_mod[id(mod)]]
+    for weapon in weapons:
+        weapon['candidates'] = _pruneDominated(weapon['candidates'], src)
+    return weapons
+
+
+def _pruneDominated(candidates, src):
+    """Drop candidates whose effective-DPS curve is dominated everywhere.
+
+    Sample each candidate's application-only multiplier (ignoring resists,
+    which are mod-independent and uniformly scale all candidates) over a
+    coarse distance grid. A candidate X is dominated if there exists Y such
+    that Y's raw_damage * multiplier(distance) >= X's at every sample.
+    """
+    if len(candidates) <= 1:
+        return candidates
+    # Sample multipliers under a neutral mid-range scenario; this captures
+    # the shape of each ammo's range envelope without depending on misc inputs.
+    sampleDistances = [0, 1000, 5000, 10000, 20000, 40000, 80000, 160000, 320000]
+    tgtSpeed = 0
+    atkSpeed = 0
+    tgtSigRadius = 125
+    sigRefMod = src.getSigRadius()  # not directly used, kept for clarity
+    del sigRefMod
+    # For each candidate, build a scalar score vector across samples.
+    scores = []
+    for snap in candidates:
+        rawTotal = snap['dmg'].total
+        vec = []
+        for d in sampleDistances:
+            if snap['kind'] == 'turret':
+                # Use only the range factor (drop tracking — angular speed is 0 here)
+                # by passing 0 atkSpeed/tgtSpeed/tgtAngle.
+                mult = _turretApplication(snap, src, src, atkSpeed, 0, d, tgtSpeed, 0, tgtSigRadius)
+            else:
+                mult = _missileApplication(snap, d, tgtSpeed, tgtSigRadius)
+            vec.append(rawTotal * mult)
+        scores.append(vec)
+    # Mark dominated
+    n = len(candidates)
+    eps = 1e-9
+    keep = [True] * n
+    for i in range(n):
+        if not keep[i]:
+            continue
+        for j in range(n):
+            if i == j or not keep[j]:
+                continue
+            # j dominates i if scores[j][k] >= scores[i][k] - eps for all k
+            # and scores[j][k] > scores[i][k] + eps for at least one k
+            dominates = True
+            strict = False
+            for k in range(len(sampleDistances)):
+                if scores[j][k] + eps < scores[i][k]:
+                    dominates = False
+                    break
+                if scores[j][k] > scores[i][k] + eps:
+                    strict = True
+            if dominates and strict:
+                keep[i] = False
+                break
+    return [c for c, k in zip(candidates, keep) if k]
+
+
+def _bestWeaponDpsAtDistance(weapon, src, tgt, atkSpeed, atkAngle, distance, tgtSpeed, tgtAngle, tgtSigRadius, profile,
+                             inLockRange):
+    if not inLockRange:
+        # Special case: FoF missiles ignore lock range
+        candidates = [c for c in weapon['candidates'] if c.get('isFoF')]
+        if not candidates:
+            return 0
+    else:
+        candidates = weapon['candidates']
+    best = 0
+    for snap in candidates:
+        if snap['kind'] == 'turret':
+            mult = _turretApplication(snap, src, tgt, atkSpeed, atkAngle, distance, tgtSpeed, tgtAngle, tgtSigRadius)
+        else:
+            mult = _missileApplication(snap, distance, tgtSpeed, tgtSigRadius)
+        scalar = _typedDmgScalar(snap['dmg'], mult, profile)
+        if scalar > best:
+            best = scalar
+    return best
+
+
+class Distance2EnvelopeDpsGetter(SmoothPointGetter):
+    _baseResolution = 50
+    _extraDepth = 2
+
+    def _getCommonData(self, miscParams, src, tgt):
+        # Snapshot per-weapon ammo candidates once. _calculatePoint reuses these
+        # for every distance step so we avoid repeated charge swaps.
+        weapons = _collectWeaponCandidates(src)
+        # Track ammo-envelope weapon IDs so we can subtract their stock contribution
+        # from the common application map below.
+        envelopeMods = {id(w['mod']) for w in weapons}
+        # Standard application path covers everything else (drones, fighters,
+        # smartbombs, doomsdays, modules without valid charges, etc.).
+        defaultSpool = eos.config.settings['globalDefaultSpoolupPercentage']
+        spoolOptions = SpoolOptions(SpoolType.SPOOL_SCALE, defaultSpool, False)
+        nonEnvelopeDmg = {}
+        for mod in src.item.activeModulesIter():
+            if id(mod) in envelopeMods:
+                continue
+            if not mod.isDealingDamage():
+                continue
+            nonEnvelopeDmg[mod] = mod.getDps(spoolOptions=spoolOptions)
+        for drone in src.item.activeDronesIter():
+            if not drone.isDealingDamage():
+                continue
+            nonEnvelopeDmg[drone] = drone.getDps()
+        for fighter in src.item.activeFightersIter():
+            if not fighter.isDealingDamage():
+                continue
+            for effectID, effectDps in fighter.getDpsPerEffect().items():
+                nonEnvelopeDmg[(fighter, effectID)] = effectDps
+        return {'weapons': weapons, 'nonEnvelopeDmg': nonEnvelopeDmg, 'tgtResists': tgt.getResists(),
+            'tgtFullHp': tgt.getFullHp()}
+
+    def _calculatePoint(self, x, miscParams, src, tgt, commonData):
+        distance = x
+        tgtSpeed = miscParams['tgtSpeed']
+        tgtSigRadius = miscParams.get('tgtSigRad', tgt.getSigRadius())
+        atkSpeed = miscParams.get('atkSpeed', 0) or 0
+        atkAngle = miscParams.get('atkAngle', 0) or 0
+        tgtAngle = miscParams.get('tgtAngle', 0) or 0
+        profile = _buildResistProfile(commonData['tgtResists'], commonData['tgtFullHp'])
+        inLockRange = checkLockRange(src=src, distance=distance)
+
+        total = 0
+        # Sum optimum-ammo contribution for each ammo-bearing weapon
+        for weapon in commonData['weapons']:
+            total += _bestWeaponDpsAtDistance(weapon=weapon, src=src, tgt=tgt, atkSpeed=atkSpeed, atkAngle=atkAngle,
+                distance=distance, tgtSpeed=tgtSpeed, tgtAngle=tgtAngle, tgtSigRadius=tgtSigRadius, profile=profile,
+                inLockRange=inLockRange)
+
+        # Add fixed-ammo contributors (drones, fighters, smartbombs, etc.) using
+        # the standard application math from fitDamageStats.
+        if commonData['nonEnvelopeDmg']:
+            applicationMap = getApplicationPerKey(src=src, tgt=tgt, atkSpeed=atkSpeed, atkAngle=atkAngle,
+                distance=distance, tgtSpeed=tgtSpeed, tgtAngle=tgtAngle, tgtSigRadius=tgtSigRadius)
+            for key, dmgTyped in commonData['nonEnvelopeDmg'].items():
+                mult = applicationMap.get(key, 0)
+                total += _typedDmgScalar(dmgTyped, mult, profile)
+        return total
diff --git a/graphs/data/fitDamageEnvelope/graph.py b/graphs/data/fitDamageEnvelope/graph.py
