quick_check: moller_analysis/h_moller_x, h_moller_y, h_moller_z are always empty (FillMollerXY/FillMollerZ never called)

[Chao1009]

Where

• Fill and getter methods: https://github.com/JeffersonLab/prad2evviewer/blob/d21ae9f/analysis/include/PhysicsTools.h#L176-L182
• Histogram booking (h_moller_x, h_moller_y, h_moller_z): https://github.com/JeffersonLab/prad2evviewer/blob/d21ae9f/analysis/src/PhysicsTools.cpp#L55-L63
• Written to the output by quick_check: https://github.com/JeffersonLab/prad2evviewer/blob/d21ae9f/analysis/tools/quick_check.cpp#L1333-L1338
• Per-thread merge only handles the phi-diff histogram: https://github.com/JeffersonLab/prad2evviewer/blob/d21ae9f/analysis/tools/quick_check.cpp#L1176
• HyCal Moller selection, which computes center and vertex but fills only ee_center_*_hc / ee_vertex_z_hc: https://github.com/JeffersonLab/prad2evviewer/blob/d21ae9f/analysis/tools/quick_check.cpp#L474-L500
• Read by the replay viewer: https://github.com/JeffersonLab/prad2evviewer/blob/d21ae9f/scripts/replay_viewer.py#L284-L294 and https://github.com/JeffersonLab/prad2evviewer/blob/d21ae9f/scripts/replay_viewer.py#L2626-L2634

What happens
Nothing in the tree calls PhysicsTools::FillMollerXY or PhysicsTools::FillMollerZ. Even so, prad2ana_quick_check writes moller_analysis/h_moller_x, h_moller_y and h_moller_z into every output file. All three histograms have zero entries on every run.

scripts/replay_viewer.py loads exactly these histograms into the "Moller Z vertex", "Moller X center" and "Moller Y center" panels of its Moller Analysis tab, and those panels have auto_cb_fit enabled. The panels are therefore always empty. moller_analysis/h_moller_phi_diff is the only filled histogram in that directory.

The two calls were removed in d17eead ("add a strict selection for e-p and e-e events yiled"). That commit moved the fits to merged->h_ee_center_x/y and h_ee_vertex_z but kept writing the PhysicsTools histograms.

Evidence

$ git grep -n -E "FillMollerXY|FillMollerZ" d21ae9f -- .
d21ae9f:analysis/include/PhysicsTools.h:176:    void FillMollerXY(float x, float y) { ... }
d21ae9f:analysis/include/PhysicsTools.h:177:    void FillMollerZ(float z) { ... }
d21ae9f:docs/PRAD2ANA_API.md:238:void  FillMollerXY    (float x, float y);
d21ae9f:docs/PRAD2ANA_API.md:239:void  FillMollerZ     (float z);

// quick_check.cpp:1334-1338
outfile.mkdir("moller_analysis"); outfile.cd("moller_analysis");
if (physics.GetMollerPhiDiffHist()) physics.GetMollerPhiDiffHist()->Write();
if (physics.GetMollerXHist()) physics.GetMollerXHist()->Write();   // never filled
if (physics.GetMollerYHist()) physics.GetMollerYHist()->Write();   // never filled
if (physics.GetMollerZHist()) physics.GetMollerZHist()->Write();   // never filled

To reproduce, run prad2ana_quick_check on any recon file and open moller_analysis/h_moller_x (or _y, _z). GetEntries() returns 0.

Suggested fix
There are two options.

1. Fill the histograms. In the HyCal Moller block (quick_check.cpp L489-L500):

   • Call physics.FillMollerXY(center[0], center[1]) next to each h_ee_center_*_hc fill. This covers both center and center2.
   • Call physics.FillMollerZ(vertex) next to h_ee_vertex_z_hc->Fill(vertex).
   • Add GetMollerXHist(), GetMollerYHist() and GetMollerZHist() ->Add(...) lines to mergeResult next to L1176. Without them the per-thread fills are lost in the merge.

   The binning differs from the _hc histograms. h_moller_x and h_moller_y have 100 bins over ±10 mm, and h_moller_z has 1000 bins over 5000–8000 mm. The _hc histograms have 800 bins over ±20 mm and 8000 bins over 5000–9000 mm.

2. Remove the duplicates. Drop moller_x_, moller_y_, moller_z_ and their Fill and Get methods from PhysicsTools and docs/PRAD2ANA_API.md. Write the existing ee_center_x_hc, ee_center_y_hc and ee_vertex_z_hc histograms into moller_analysis/ instead, and update the keys in scripts/replay_viewer.py.

Status
Still present on the dedup-refactor branch. Nothing calls FillMollerXY or FillMollerZ, and quick_check.cpp still writes the three histograms.