PyReconstruct User Guide¶

PyReconstruct is a desktop application for tracing, annotating, and 3D-reconstructing serial-section and volume electron-microscopy (EM) data. This guide covers installing the app, opening and building a series, the tracing tools, the data lists, alignment, 3D reconstruction, and backups.

This guide describes the application as it currently behaves in this distribution. It complements the in-app help (Help ▸ Online resources ▸ PyReconstruct user guide, which opens the lab wiki at The University of Texas at Austin) and the in-app Help ▸ Shortcuts list.

About screenshots. The screenshots and clips show PyReconstruct on macOS with a small test series. They are shot from PyReconstruct itself, and shot again when PyReconstruct changes.

Contributions of screenshots are welcome. See CONTRIBUTING.md.

Shortcuts. Key combinations shown here are the defaults. Almost all of them are rebindable under Help ▸ Shortcuts list (press ?). A few menu shortcuts are fixed: PgUp/PgDown (section navigation) and ? (the shortcuts list itself). This guide writes Ctrl and Alt; on macOS press Cmd and Option. Keyboard shortcuts lists both forms.

1. Installing PyReconstruct¶

There are two ways to install PyReconstruct: a one-click installer (recommended for most users) or a from-source install with pip (for developers and other platforms).

One-click installers (Windows, macOS, and Linux)¶

Download the build for your platform from the current stable release, v1.25.0. Windows, macOS, and the Linux AppImage include Python. Each link below downloads that platform's installer directly; every build, its SHA-256 checksum, and past versions are on the Releases page.

Windows: PyReconstruct-1.25.0-Windows-x86_64-Setup.exe. This is a per-user installer (no administrator rights required). Builds are unsigned for now, so Windows SmartScreen may warn that the publisher is unknown; choose More info ▸ Run anyway. Re-running a newer installer upgrades the existing installation in place.
macOS: open the .dmg and drag PyReconstruct onto the Applications shortcut. PyReconstruct is signed and notarized by Apple, so the first launch shows the usual question for an app downloaded from the internet. Click Open.

Versions before 1.24.0 were not signed. To open one of those, clear the quarantine flag once in Terminal:

xattr -dr com.apple.quarantine /Applications/PyReconstruct.app

The PyReconstruct installer window on macOS: the PyReconstruct app icon on the left, an arrow, and the Applications folder shortcut on the right.

Both macOS builds are native: PyReconstruct-1.25.0-macOS-arm64.dmg for Apple Silicon and PyReconstruct-1.25.0-macOS-x86_64.dmg for an Intel Mac. The in-app updater serves each Mac its matching architecture.

Linux: install the AppImage with one command:

curl -fsSL https://pyreconstruct.org/install.sh | bash

It needs no root and no Python. It adds a pyreconstruct command and a menu entry. Re-run it to update. For PyReconstruct Dev, end the command with bash -s -- --dev. To uninstall, end it with bash -s -- --uninstall. You can also download PyReconstruct-1.25.0-linux-x86_64.AppImage and run it directly, or install from source with a system Python 3.11, 3.12 or 3.13:

curl -fsSL https://pyreconstruct.org/install-from-source.sh | bash

If FUSE is missing, PyReconstruct still starts, only more slowly, and the installer names the package that speeds it up.

From source (other platforms and developers)¶

PyReconstruct requires Python 3.11, 3.12 or 3.13. Its native dependencies are tested on those versions (the project pins >=3.11,<3.14). Your system python3 may be newer, and installing against it will fail; the steps below get you 3.11 without changing your system Python (use --python 3.12 or --python 3.13 in place of --python 3.11 for those). The recommended tool is uv, which downloads Python 3.11 for you. Install it with curl -LsSf https://astral.sh/uv/install.sh | sh (or brew install uv).

For a quick, non-editable install into a fresh 3.11 environment:

uv venv --python 3.11 && source .venv/bin/activate   # Windows: .venv\Scripts\activate
uv pip install git+https://github.com/dustenhubbard/PyReconstruct
PyReconstruct

pip pulls in the runtime dependencies (PySide6, VTK, vedo, NumPy, SciPy, scikit-image, shapely, trimesh, zarr, and others). If you already have python3.11 on PATH, plain venv works instead of uv: python3.11 -m venv .venv && source .venv/bin/activate, then pip install git+….

To track the latest unreleased code on main (what the retired in-app "Developer" channel used to offer), clone the repository and let uv build the environment from the committed uv.lock. This is the canonical developer setup: uv sync reads the checkout's .python-version (3.11), provisions the interpreter, and installs the exact pinned dependency set. To use 3.12 or 3.13 instead, set UV_PYTHON=3.12 (or 3.13) in your shell before uv sync and uv run.

git clone https://github.com/dustenhubbard/PyReconstruct
cd PyReconstruct
uv sync                            # creates .venv from uv.lock (exact pinned deps)
uv run PyReconstruct               # launch

To move to the newest main later: git pull, then uv run PyReconstruct. There is no manual reinstall step, because uv run re-syncs .venv to the lockfile automatically.

If you prefer a plain venv and already have python3.11 on PATH, an editable install works too, though it resolves dependencies fresh rather than from uv.lock: python3.11 -m venv .venv && source .venv/bin/activate, then pip install -e . (rerun only when pyproject.toml changes; a bare git pull otherwise suffices).

For a full development setup (test suite, dev tooling, the parallel conda pyrecon_dev environment, code layout), see CONTRIBUTING.md and DEV_UV.md.

Launching the app¶

From an installer: launch PyReconstruct like any other installed application.
From a source install: run PyReconstruct on the command line.

The command accepts a few optional flags, e.g. PyReconstruct -f path/to/series.jser to open a series directly on launch. (Run PyReconstruct --help for the full list.)

2. Keeping PyReconstruct up to date¶

The frozen Windows and macOS one-click builds can update themselves from within the app. The updater downloads the new build from GitHub Releases and verifies it against a published SHA-256 checksum before installing. If the checksum can't be reached or doesn't match, nothing is installed. (On Linux, re-run the install command to update; see Installing PyReconstruct.)

Update channels¶

Two apps install side by side, each following its own update channel:

PyReconstruct: Stable. Stable builds tagged vX.Y.Z, for everyday work.
PyReconstruct Dev: Nightly. Early features with less testing, tagged vX.Y.Z.devYYYYMMDD (for example v1.24.0.dev20260927). A new build is scheduled daily when there are new commits.

The channel follows the app and cannot be changed in a series. Use the Help menu's download link to install the other app. Each app only offers updates from its own channel, so updating Dev keeps Stable in place. Nightlies skip the What's New popup; their release notes list the changes since the previous nightly.

Checking for updates¶

Choose Help ▸ Check for updates…. On an installed build, PyReconstruct checks GitHub Releases on the app's channel (off the main thread, so the UI stays responsive). If a newer build is found, an update dialog opens showing the new version, the download size, and a link to the release notes, with a Download & Install button. After verifying the download, PyReconstruct asks to save any unsaved work, closes, and opens the installer. Complete installation, then launch the app again. Canceling the close also cancels the installation.

If you are already current, you'll see a brief "You're already up to date" message; if no installer exists for your platform on that channel, it tells you so.

The update dialog offering 9.9.9, an example version that does not exist: You have 1.25.0, channel Stable, download 178.7 MB, a Release notes for 9.9.9 link, and Later and Download & Install buttons.

The picture shows an example offer. Version 9.9.9 does not exist.

Checking automatically on startup¶

Under Help ▸ Automatically check for updates you can turn the startup check on or off. It is on by default. When enabled, an installed build does a quiet background check at most once per day; if an upgrade is available, Update available and the version appear on the right of the status bar. The notice stays, across launches too, until you click it; clicking it opens the update just like Help ▸ Check for updates…. Turning the automatic check off also removes the notice.

The Update available notice on the right of the status bar, for the example version 9.9.9.

Source / `pip` installs¶

For a from-source install, Help ▸ Check for updates… instead reinstalls PyReconstruct from a chosen Git branch with pip and then restarts. A prompt asks for the branch, starting with your saved choice or main. The channel/installer machinery above applies only to packaged builds.

The same reinstall is available from the command line: PyReconstruct --update (reinstall the current branch) or PyReconstruct --switch <branch> (change branch, then reinstall). If you installed editable (pip install -e .), a plain git pull is usually all you need. The working tree is what runs. A source install on main is the supported way to follow every commit now that the in-app Developer channel has been removed.

3. Core concepts¶

PyReconstruct is organized around serial-section microscopy: a block of tissue is cut into an ordered stack of thin sections, each imaged, and structures are traced through the stack.

Series. A complete project: an ordered set of sections plus series-wide settings. You work on one series at a time. A series is stored as a single .jser file (see below).
Section: one cross-section in the stack. Each section has an index, a background image, a magnification (mag, in µm per image pixel), a thickness (in µm), brightness/contrast, and its own set of traces.
Trace: a single connected shape or curve drawn on one section. A trace can be closed (an outline enclosing area, e.g. a cell cross-section) or open (a curve with length but no enclosed area, e.g. a process or a measurement line).
Contour: all traces with the same name on a single section.
Object: all traces with the same name throughout the whole series. Objects are what you reconstruct in 3D and measure in the object list.
Z-trace: a single curve that runs across sections (its points carry a section index), used to measure distances through the volume.
Flag: a labeled, optionally colored marker placed at a point on a section, with comments and a resolved/unresolved state. Flags are useful for to-do notes, questions, and review.
Transform. A section's affine transform, six numbers a b c d e f: a b d e give rotation/shear/scale and c f give x/y translation. The same transform is applied to the section's image and its traces, so traces stay fixed to the image. Moving the alignment moves image and traces together.
Alignment: a named set of per-section transforms across the whole series. A series can hold several alignments and you can switch between them. (See Alignment.)
Host / traveler. An optional parent/child relationship between objects: an object can be hosted by ("ride on") another object. Used to keep related structures organized and to group them in the 3D scene.

The `.jser` file¶

A series is saved as a single .jser file. It is a JSON document with the series-level settings, the per-section data (including each section's image filename, magnification, and thickness), and the edit log. The .jser does not contain the images themselves: it stores image filenames, and the actual image files live in the series' image source directory. While a series is open, PyReconstruct unpacks it into a hidden working folder next to the .jser and repacks on save.

Images and Zarr¶

Supported image formats for sections are JPG/JPEG, PNG, TIF/TIFF, and BMP.

For large images, PyReconstruct can use multiscale Zarr image stores, which load only the parts of the image needed for the current view (rather than the whole image on every section change). You can convert a series' images to scaled Zarr from Series ▸ Images ▸ Convert to scaled images (see Opening and creating a series).

4. Opening and creating a series¶

On first launch¶

With no file specified, PyReconstruct opens a welcome series, a small read-only demo you can use to explore the interface. The welcome series cannot be saved or backed up; create or open a real series to begin work. A username is resolved automatically and recorded against the edits you make. PyReconstruct no longer prompts for it on launch; change it anytime under File ▸ Change username….

The main window on first launch, showing the welcome series: a PyReconstruct logo image on the field, with the tool palette on the right and the trace palette at the bottom.

Opening an existing series¶

File ▸ Open series… (Ctrl+O) opens a .jser file. Recently opened series are listed under File ▸ Open recent series, most recently opened first (by when you opened them, not by file date); the list holds ten entries, drops paths that no longer exist, and can be emptied with Clear recents at the bottom of that submenu. File ▸ Close series returns to the welcome series (it does not quit the app).

Clip: over the welcome series, the pointer opens the File menu and clicks Open series..., and the shapes1 test series opens on section 0. — **File ▸ Open series…**, then pick the `.jser` file. The file dialog is left out here.

A small test series open on section 0: a star, a circle, a triangle, a diamond and a rectangle traced over a gray image.

If a series' image files can't be found when it opens, PyReconstruct asks whether you'd like to locate them and lets you pick the image folder (you can also do this later via Series ▸ Images ▸ Find/change image directory).

Recovering an unsaved session. If a series was left open or not saved cleanly (for example after a crash), reopening it detects the leftover working folder and offers to recover the unsaved session. If the same series appears to be open in another window (it was touched within the last few seconds), PyReconstruct shows "Series In Use" and declines to open a second copy.

Creating a new series from images¶

File ▸ New series ▸ From images… (Ctrl+N) walks you through:

Select images: choose the section image files (JPG/JPEG, PNG, TIF/TIFF, BMP). The files are sorted alphanumerically, and that order becomes the section order (section 0, 1, 2, …), so name your images so they sort correctly.
Series name: used to name the series.
Image calibration (µm/px): microns per image pixel. Default 0.00254.
Section thickness (µm): default 0.05. Thickness feeds the volume and flat-area calculations in the object list.

The new series opens fitted to the first image; you are then prompted to choose where to save the .jser.

The New Series prompt "Enter image calibration (µm/px)", with the default 0.002540, and Cancel and OK buttons.

Other ways to create a series¶

Also under File ▸ New series:

From scaled images… builds from an existing scaled Zarr directory (expects a scale_1 folder), then runs the same name/calibration/thickness prompts.
From legacy .ser… creates a series from legacy Reconstruct XML .ser files.
From neuroglancer zarr… creates a series from a Neuroglancer Zarr.

Image directory and Zarr conversion¶

Under Series ▸ Images:

Find/change image directory points the series at the folder containing its images. The chosen directory is remembered with the series.
Convert to scaled images converts the series' images into a scaled (.zarr) store for faster loading of large images.
Update image scales adds or refreshes scale levels on an existing Zarr.

When a Zarr names its images differently¶

A Zarr made from a series' images may name them differently from the series, for example IMG_000.tif in the series and 000_IMG_grid000.tif in the Zarr. The series then shows no images, even with the right Zarr chosen.

Choose the Zarr in Series ▸ Images ▸ Find/change image directory. If none of the series' image names is in the Zarr, PyReconstruct tries to pair each section with one Zarr image by the number in its image name:

It uses the first or the last number in each name. Zero padding does not count, so 001 and 1 are the same number.
Each number must appear once on each side, and both sides must have the same numbers. An extra image, a missing image or an offset (for example images numbered 1 to 100 against 0 to 99) is refused.
If the first and the last number both fit but pair the images differently, the match is refused, since which one is meant is not clear.
The pairing uses the number in the series' own image name, not the section number. After you insert, remove or reorder sections, the two can differ.

If the names cannot be paired, a message says why and no names change. If even one series image name is already in the Zarr, PyReconstruct treats the rest as missing images and offers no pairing.

If the names pair, a Match Image Names window lists each section with its current image name and the Zarr name it would get. It draws a few sections' traces over the proposed images so you can check the fit by eye. Sections whose image number is not their section number are flagged and drawn first.

Use these names gives each section its Zarr image name. The series file changes when you save.
Cancel changes no names. The series still points at the chosen Zarr, so its images still do not show.

Matching the names only reads the Zarr. Nothing in it is renamed, moved or written, so other series that use the same Zarr are not affected. If the Zarr has no scale levels yet, PyReconstruct may then ask whether to add them. That is a separate step: Yes writes the new scale levels into the Zarr, and No leaves it as it is.

The main window shows the current section's image with its traces drawn on top (the field). Floating overlays sit on top of the field: the tool palette (the column of mode buttons, top-right by default), the trace palette (the row of preset traces, bottom-center), section-increment buttons, brightness/contrast sliders, and a scale bar. Each overlay group can be dragged to reposition it, and View ▸ Reset palette position restores the defaults.

The main window with its parts labeled: the field, the tool palette at the top right, the brightness and contrast sliders, the section buttons at the bottom right, the trace palette at the bottom center, and the scale bar at the bottom left.

Moving between sections¶

Page Up / Page Down: next / previous section.
Mouse wheel over the field: scroll up for the next section, down for the previous (works in any tool mode).
Go to section (Ctrl+G): jump to a section number.
The section-increment buttons (▲/▼) in the corner do the same as PgUp/PgDown (handy on tablets).

Clip: the pointer clicks the up arrow button at the bottom right and section 1 shows, then clicks the down arrow and section 0 is back. — **Click ▲ for the next section and ▼ for the previous one.**

Panning and zooming (in any tool)¶

Middle-click and drag: pan the view.
Ctrl + mouse wheel: zoom about the cursor.
Home: fit the view to the image. View ▸ View magnification… sets an exact zoom (useful for consistent figure scale).

The dedicated Pan/Zoom tool (below) is an alternative when a middle mouse button isn't available.

Brightness and contrast¶

Adjust the current section with the corner sliders, or with the keyboard: brightness = / -, contrast ] / [ (also under Edit ▸ Brightness/contrast). Named brightness/contrast profiles are available under Series ▸ Brightness/contrast profiles….

Showing and hiding interface elements¶

Under View ▸ Palette ▸ Visibility you can independently toggle the trace palette, the section-increment buttons, the brightness/contrast sliders, and the scale bar. View ▸ Left handed swaps the tracing-pencil cursor side.

Theme¶

View ▸ Change theme offers Default and Dark. (The theme can also be set under Series ▸ Options.)

The main window in the Dark theme: the menu bar, status bar and palette buttons turn dark gray.

6. The tool palette¶

The tool palette is the column of mode buttons (top-right by default). Click a button (or press its shortcut) to switch the active tool. Right-clicking a tool button opens that tool's settings (where it has any).

In every tool except Pan/Zoom, right-clicking the field opens a context menu (for the selected traces, or for the field). The mouse wheel changes sections, and middle-click pans, regardless of the active tool.

The tool palette column, each button labeled with its tool and default shortcut: Pointer (P), Pan/Zoom (Z), Knife (K), Scissors (Shift+K), Closed Trace (C), Open Trace (O), Stamp (S), Grid (G), Flag (F), Host (Q), and Ztool, which has no shortcut.

The palette has eleven tools, top to bottom:

Tool	Key
Pointer	`P`
Pan/Zoom	`Z`
Knife	`K`
Scissors	`Shift+K`
Closed Trace	`C`
Open Trace	`O`
Stamp	`S`
Grid	`G`
Flag	`F`
Host	`Q`
Ztool	(none)

Pointer (`P`)¶

Selects and moves traces, z-trace points, and flags.

Click a trace to select/deselect it.
Click and drag on empty space to draw a selection region. By default this is a freehand lasso that selects only fully-enclosed traces; right-click the Pointer button to switch the region shape (Rectangle/Lasso) and whether it selects all touched traces or only completely encircled ones.
Click and drag a selected trace to move all selected items together.

The Pointer settings dialog: Shape (Rectangle or Lasso), Select (All touched traces or Only completed encircled traces), and Display closest field item.

Pan/Zoom (`Z`)¶

Drag with the left button to pan.
Drag up/down with the right button to zoom.

(Pan and zoom are also available in any tool via middle-click and Ctrl+wheel.)

Knife (`K`)¶

Slices a selected trace along a freehand line.

Select the trace(s) to cut first (they must share one object name and be all open or all closed), then drag the knife across them.
Resulting pieces smaller than a threshold percentage of the original are discarded. Right-click the Knife button to set % original trace (default 1.0) and to enable Smooth cuts.
A press of another mouse button part way through a cut is ignored, so a drawing tablet's barrel button cannot cancel the stroke or open the field menu over the object. Clear Ignore the other mouse buttons in the same right-click dialog to get the older behavior, where the second press abandons the cut and opens the menu.
A trace whose outline crosses itself cannot be cut. The knife says so and leaves the object alone; smooth or redraw the trace and cut again.

The Knife settings dialog: % original trace set to 1.0, Smooth cuts, and Ignore the other mouse buttons, which is checked.

Scissors (`Shift+K`)¶

Re-traces an existing trace. Click a trace and the tool reopens it as a live line starting at the clicked point, so you can re-draw it; right-click to finish. Useful for fixing part of a contour without redrawing the whole thing.

Closed Trace (`C`) and Open Trace (`O`)¶

Draw new traces with the current trace-palette preset's name and color. Closed Trace makes outlines that enclose area; Open Trace makes curves with length only.

Clip: the pointer clicks the Closed Trace button, drags a loop beside the star and lets go, and the new orange trace stays selected. — **Click Closed Trace (or press `C`), then drag a loop and let go.** The trace takes the selected preset's name and color.

Right-click either button to choose the drawing mode:

Scribble: hold and drag to draw freehand; release to finish.
Poly: click to place each vertex; right-click to finish; Backspace removes the last point.
Combo (default): a click starts a poly/clicked line, however long the button is held, if the pointer stays still; click-and-drag scribbles.

The Closed Trace settings dialog: Mode (Scribble, Poly or Combo), Closed Trace Shape (Trace, Rectangle or Ellipse), Automatically merge overlapping traces, and Apply rolling average while scribbling.

For Closed Trace you can also pick a fixed shape (freehand Trace, Rectangle, or Ellipse; drag to size) and optionally auto-merge new traces into selected same-named traces, or apply a rolling-average smooth while scribbling.

A new closed trace, drawn freehand in orange beside the star with the Closed Trace tool.

Stamp (`S`)¶

Places a copy of the current trace-palette preset's shape at a fixed size.

Click to drop one stamp centered on the cursor at the preset's radius.
Click and drag to set the radius for that stamp (diameter = drag distance).

The stamp shape and its radius come from the selected trace-palette button (set them by right-clicking that button; see The trace palette).

Grid (`G`)¶

Replicates the current preset into a rectangular array, useful for stereology sampling. Click to place the grid. Right-click the Grid button to set the element size, spacing, and number of columns/rows. With Sampling frame enabled (the default), each cell is drawn as a counting frame (a red exclusion line and a green inclusion line) rather than a copy of the preset.

The Grid settings dialog: element size, distance and number in X and Y, and Sampling frame, which is checked.

Flag (`F`)¶

Drops a labeled marker (a ⚑ glyph) at the clicked point. Click to place a flag and fill in its name, color, and an optional comment. Right-click the Flag button to set the default name/color, the flag size, and which flags are shown: All flags, Only unresolved flags (the default), or No flags.

Flags are selected and moved with the Pointer tool; right-clicking a flag opens its dialog, where you can edit its name/color, add comments, and mark it Resolved. All flags in the series are listed in the Flag list.

Host (`Q`)¶

Sets host (parent) relationships between objects. Click a first trace (the object that will be hosted), then click a second trace (its host); a line is drawn between them while you choose. Right-click to cancel. An object can't host itself, and two objects can't host each other. Hosts also appear as a column in the object list and can be set from the object list's context menu.

Z-trace tool¶

Creates a z-trace, a curve through the stack. The palette names this tool Ztool. Select the tool, then click points (changing sections between clicks as needed with the mouse wheel); right-click to finish. The new z-trace takes the current preset's name and color. (This tool is selected from the palette; it has no default keyboard shortcut.)

7. The trace palette¶

The trace palette is the row of preset traces at the bottom of the window. The selected preset supplies the name, color, tags, shape, and stamp radius for new traces you draw, stamp, or grid. The selected preset's name is shown (in its color) near the palette.

The trace palette: two rows of ten preset buttons, with the selected preset's name, circle1, above them in orange.

The default palette has 20 presets, arranged as two rows of ten.
Select a preset with the number keys: 1–9, 0 for the first row, and Shift+1–Shift+9, Shift+0 for the second.
Right-click a preset (or press Ctrl+its number) to edit its attributes: name, color, fill mode, tags, the trace shape, and the stamp radius (microns) used by the Stamp and Grid tools.

The small buttons beside the palette let you bulk-increment numbered preset names (+/−, applied to all presets or only the active one), open Modify all palettes, and show help. Preset names containing a number pattern can also auto-increment as you place traces, which is handy for numbering a sequence of objects.

You can reset the palette (Series ▸ Trace palette ▸ Reset current palette) and export/import palettes as CSV (Series ▸ Trace palette ▸ Export as CSV… / Import from CSV…). Edit ▸ Paste attributes to palette (Shift+G) sets the current preset from a copied trace.

The dialog a right-click on a preset opens: Name, Color, Shape, Fill, Stamp radius (microns), Trace Tags and Object Groups.

8. Working with traces¶

Most trace editing is on the field context menu (right-click selected traces), with keyboard shortcuts for the common actions. Selected-trace actions affect all selected traces. The top of that menu carries a single Edit ... attributes... shortcut that follows your selection: it reads Edit trace attributes... when traces are selected and Edit z-trace attributes... when z-traces are, so the most-used edit is one click away above the per-entity submenus.

Edit attributes (Ctrl+E): change name, color, tags, and fill mode.
Merge traces (Shift+M): merge the exteriors of selected traces (they must share a name).
Hide (Shift+H) / Unhide all (Ctrl+U). Hidden traces can't be edited until unhidden.
Make negative / positive. Negative traces subtract from (cut into) the area of same-named traces, e.g. to carve a hole; this matters for area and 3D volume.
Cut / Copy / Paste (Ctrl+X / Ctrl+C / Ctrl+V) and Paste attributes (Ctrl+Shift+V; the older Ctrl+B still works), which copies a trace's name/color/tags onto selected traces.
Delete (Delete or Backspace): delete the selected traces.

On the field (right-click empty space, or use the shortcuts):

Select all (Ctrl+A) / Deselect (Ctrl+D).
The View submenu items are checkboxes that show their current on/off state; each also has a keyboard shortcut that toggles it:
Hide trace layer (H) and Show all traces (ignore hidden) (A) temporarily hide or show every trace regardless of individual hidden state (the field border turns red when all are force-hidden, green when all are force-shown). These differ from per-trace Hide/Unhide.
Section blend (Space) blends the current and last-viewed section, to compare them.
Hide image (I), Focus mode (X).

In Focus mode, a Ctrl-click (Cmd-click on macOS) on a trace edits which object it belongs to: clicking a trace of the focused object splits that trace out into a new <object>_split contour, and clicking a trace of any other object incorporates it into the focused one. Both are undoable. This was Shift-click before 1.21.0, and Shift no longer performs the edit. Choose another modifier, or none, under Help ▸ Shortcuts list. Clicking a trace without a modifier does nothing in focus mode. The ordinary click-to-select is suppressed outright while focus mode is on, so an unmodified click neither selects another object's trace nor deselects one of the focused object's own.

Undo (Ctrl+Z) / Redo (Ctrl+Y) cover actions on the field and actions taken through the lists, including series-wide edits such as renaming, recoloring, or changing the radius or shape of an object across every section. When an action touched more than one section, undo asks whether to revert all of them or only the current section. Three actions are genuinely not undoable and warn before running: deleting sections, reordering sections, and optimizing brightness/contrast.

Find a contour on the current section with Section ▸ Find contour… (Shift+F); jump to an object's first contour anywhere in the series with Series ▸ Find first object contour… (Ctrl+F).

9. Data lists¶

PyReconstruct provides five list/table windows, plus a series history window, under the Lists menu. Each list opens as a dockable panel (multiple instances are allowed), and several can be open at once.

List	Open with	Shows
Object list	`Ctrl+Shift+O`	every object in the series, with quantities and attributes
Object list of selected traces	(no shortcut)	the objects with a trace selected in the field, with the same columns
Trace list	`Ctrl+Shift+T`	traces on the current section
Section list	`Ctrl+Shift+S`	every section
Z-trace list	`Ctrl+Alt+Z`	every z-trace
Flag list	`Ctrl+Shift+F`	flags across the series
Series history	(no shortcut)	the full edit log

Features shared by the lists¶

Set columns… lets you choose which columns are shown and reorder them.
Export… writes the displayed columns/rows to CSV.
Refresh reloads from the series data (the object list also auto-refreshes as you trace; the trace list tracks the current section and has no separate Refresh).
Filters. Most lists filter by a regex on the name (where # is shorthand for a digit), and by group and/or tag where applicable.
Ctrl+C copies selected cells. Delete/Backspace deletes the selected rows' underlying data. Right-click a selection for that list's context menu.
Every list's context menu offers Invert selection (select the displayed rows that aren't selected, and vice versa; it only ever acts on rows currently shown, so an active filter is respected) and Copy entity values (named for the list: Copy object values, Copy trace values, etc.; the same as Ctrl+C).

The Object list docked to the left of the field, with its List, Selection, Filter, Find, Groups and Columns menus, and one row per object.

Object list¶

Double-click an object to jump to its first occurrence; Shift+double-click adds it to the 3D scene. Columns include Name, Start/End (first/last section), Count, Flat area, Volume, Radius, Host/Superhosts, Groups, Trace tags, Locked, Last user, Comment, Configuration (open/closed/mixed), an optional Curate state, and any custom categorical columns you define.

Flat area = the summed closed-trace areas plus open-trace lengths × section thickness; Volume = summed closed-trace areas × section thickness.
The context menu edits the object across the whole series: edit attributes, set comments/hosts/groups/alignment, lock/unlock, create a copy, edit radius/shape, smooth, split into individual objects, set curation, manage the object in 3D (add/remove/export meshes/export quantitative data/edit 3D settings), create a z-trace from it, view history, and delete.
Filters: regex, group, tag, curation, configuration, and host filters. Filter ▸ Selected traces only limits the list to the objects with a trace selected in the field, and the rows update when the selection changes.
Find ▸ First / Last jump to the object's first/last section.
The Curate columns can be toggled across all object lists with View ▸ Toggle curation in object lists (Ctrl+Shift+C).
Series ▸ Restore object curation status from log recovers curation statuses the series has lost. Every time you mark an object Curated or Needs curation the series records it twice: on the object, and as an entry in the series log. This action reads the log (including entries offloaded with Series ▸ Log ▸ Offload log history…), takes each object's most recent curation entry, and writes that status back onto the object. It only fills gaps: an object already marked Curated is left exactly as it is, and no status is ever cleared or downgraded. A restored Needs curation keeps who it was assigned to, provided the marking was made in a version that records the assignee in the log. Use it if curation you know you did is no longer showing, for example after recovering an unsaved session. It is one undoable step (Ctrl+Z).

Trace list¶

Lists traces on the current section (it follows you as you change sections). Columns include Name, Index, Tags, Hidden, Closed, Length, Area, Radius, optional Centroid and Feret diameters. List ▸ Export all traces in series… writes every trace in the series to CSV. The context menu can edit/smooth/merge traces, set open/closed, make negative/positive, hide, edit shape/radius, create a flag on a trace, and delete.

The Trace list docked to the left of the field, with a row for each trace on the current section.

Section list¶

Lists all sections. Columns: Section (number; calibration-grid sections are marked), Thickness, Locked, Brightness, Contrast, Image Source. The context menu locks/unlocks sections, sets brightness/contrast (set, increment, match to the in-view section, or optimize), edits thickness, edits the image source, inserts a section above/below, copies, and deletes. Modify ▸ Section image sources… bulk-renames image sources, and Reorder sections renumbers them sequentially.

The Section list docked to the left of the field, with a row for each section: thickness, locked, brightness and contrast.

Z-trace list¶

Lists all z-traces. Columns: Name, Start, End, Distance (length), Groups, Alignment. The context menu edits attributes, smooths, manages the z-trace in 3D and in groups, changes its alignment, and deletes.

The Z-trace list docked to the left of the field, with the name, start, end and distance of each z-trace.

Flag list¶

Lists flags across the series; resolved flags are hidden by default. Columns: Section, Color, Flag (name), Resolved, Last Comment. Double-click jumps to the flag. The context menu opens the flag to view/edit and add comments, marks flags resolved/unresolved, filters by color, copies, deletes, or deletes all flags with a given name. Filters include showing resolved flags, a name regex, a color filter, and a comment-text filter.

Series history¶

Lists ▸ Series history opens a read-only History window (Date, Time, User, Object, Sections, Event), newest first. The object list's View history shows the same window filtered to the selected objects. Long histories can be trimmed with Series ▸ Log ▸ Offload log history…, which exports older entries to an external CSV.

Older versions could write a history row PyReconstruct cannot read back, when a name had a line break or a comma in it. From a source install, PyReconstruct --check-history path/to/series.jser lists those rows with the line and text of each. Give it a folder to check every .jser in it and its subfolders. It only reads the files and changes nothing.

10. Alignment¶

An alignment is a named set of per-section transforms. The currently active alignment determines how each section's image and traces are positioned. Switching alignments never changes the underlying images or trace coordinates, only how they are displayed.

Every series includes a special no-alignment entry (the identity transform). It can't be edited, renamed, or deleted, and you can't edit transforms while it is active. Switch to a real alignment first.

Switching and managing alignments¶

Switch the current alignment from the field's right-click menu (Series alignment submenu), which lists every alignment as a checkable item.
Alignments ▸ Edit alignments… (Ctrl+Shift+A) opens the alignment dialog, where you can create a New alignment (created as a copy of the current one), Rename, or Remove alignments. Objects can also carry a per-object alignment override (Edit alignment… in the object list).

The alignment dialog: a list of the alignments, default and no-alignment, with Remove, Rename and New buttons.

Adjusting a section's transform¶

Alignments ▸ Edit transformation (Ctrl+T): enter the six transform numbers a b c d e f directly. (Blocked on locked sections and while in no-alignment.)

The Edit section transform prompt, with the identity transform 1 0 0 0 1 0.

Keyboard nudges (when no traces are selected, these move the section's transform; otherwise they move the selected traces):
Arrow keys: translate (medium step); Ctrl+arrows = small, Shift+arrows = big. Step sizes are set in Series ▸ Options.
Ctrl+Shift+Left / Ctrl+Shift+Right: rotate about the cursor.
F1–F4 (with Shift to reverse): scale and shear in X/Y.

Assisted alignment¶

Estimate affine transform aligns the current section to the comparison ("B") section from matched traces. Select 3 or more traces of the same name on both sections (the same number on each); PyReconstruct computes the affine transform that best maps one set of centroids onto the other.
Align by correlation (Ctrl+\) automatically registers the current image to the section beneath it by image cross-correlation (a translation-only adjustment).
Propagate transform records a transform adjustment and applies it across many sections: Start propagation recording, make your adjustment, then Propagate to start / Propagate to end (or simply navigate; while recording, moving to a new unlocked section applies the recorded change). A red dot is shown on the field while recording. Locked sections are skipped.

Align by correlation, then propagate across a range¶

The Ctrl+\ correlation shift can be recorded and propagated across a range of sections, exactly like a manual transform, but you must start recording before you correlate:

Navigate to the section you want to align.
Alignments ▸ Propagate transform ▸ Start propagation recording.
Press Ctrl+\ (Align by correlation). This aligns the current section and records the shift.
Either navigate through the sections (each unlocked section you visit receives the recorded shift) or use Propagate to start / Propagate to end to apply it across a bulk range at once.
Alignments ▸ Propagate transform ▸ End propagation recording.

Order matters: starting recording resets the accumulator, so pressing Ctrl+\ before you start recording will not capture that alignment. Always start recording first. As with manual transforms, locked sections are skipped during propagation.

Importing transforms¶

Alignments ▸ Import alignments:

Another series (.jser): choose a series, then choose which of its alignments to bring over and what to call each one here. Each name starts out the same as it is in the other series; type over it to import under a different name. Importing under a name this series already uses replaces that alignment on every section, and PyReconstruct names the alignments at risk and asks before it does. Differing calibrations are flagged first, and transforms are rescaled to this series' pixel size. Sections missing from the other series get an identity transform.
.txt file. One line per section: section a b c d e f (the integer section number followed by the six transform numbers). Every section number must exist in the series; the translation terms are interpreted in pixels and scaled by the section magnification. Blank lines are skipped, and any other line that does not fit stops the import with a message that names it. A file with no transforms stops it too. The imported transforms are written to a new alignment named after the file (with the date appended), and the series switches to it.
SWiFT project: import transforms from an AlignEM-SWiFT project; you choose the scale to import. The number of transforms must match the number of sections. They go to the sections in order, so the series does not have to start at section 0. With Includes cal grid checked, the section marked as the cal grid gets the identity transform and the SWiFT stack goes to the other sections. If no section or more than one is marked, the first section gets it, and the message after the import says so. A file that is not a SWiFT project, or has no transforms at the chosen scale, stops the import with a message.

The same alignment import is also one tab of Series ▸ Import series data ▸ From another series…, which is where to go when you want alignments and traces, flags, or palettes in a single pass.

Locking sections¶

A locked section's transform can't be changed by any means (manual edit, nudges, estimate/correlate, or propagation), and locked sections are protected from brightness/contrast changes, thickness edits, image-source edits, and deletion. Lock/unlock from the Section list (the Locked column or its context menu); unlock the current section quickly with Alignments ▸ Unlock current section (Ctrl+Shift+U).

11. 3D reconstruction¶

Generating a 3D object¶

From the Object list, select one or more objects and choose 3D ▸ Add to scene from the context menu (or Shift+double-click an object). The first time, this opens the 3D scene window; if a scene is already open, the objects are added to it. Meshes are built on a background thread (you'll see a "Generating 3D…" progress indicator). Z-traces can also be added to the scene (as tubes) from the z-trace list.

Clip: the pointer opens Lists ▸ Object list, clicks ast01, Shift-clicks d01sp01 to select all four objects, right-clicks them and chooses 3D ▸ Add to scene; the 3D scene opens with a dendrite, its spine, an axon and an astroglial process. — **Lists ▸ Object list**, click the first object, **Shift**-click the last, right-click the selection, then **3D ▸ Add to scene**. The series is a made-up one: d01 a dendrite, d01sp01 its spine, ax01 an axon, ast01 astroglia.

The 3D scene window with four objects as surfaces, a star, a triangle, a rectangle and a circle, and a gray one-micron scale cube.

3D model types¶

Each object has a 3D type (set in Edit 3D settings… from the object list):

Surface (default): a smoothed surface reconstructed from the object's traces (negative traces are subtracted, e.g. to hollow out a structure).
Spheres: one sphere per trace, sized to the trace's radius (good for point-like or stamped objects).
Contours: each trace rendered as a thin slab, showing the raw cross-sections.

Z-traces render as tubes. Edit 3D settings… also sets per-object opacity.

Series-wide 3D quality is controlled in Series ▸ Options (the 3D section): an XY Resolution slider (less detail/faster ↔ more detail/slower), the 3D smoothing method (Humphrey is recommended; the others are Mutable Diffusion Laplacian, Taubin, and None), the smoothing iterations, and the screenshot resolution (dpi).

Navigating the 3D scene¶

Left-drag rotates; middle-drag pans; right-drag zooms. (Ctrl-drag also rotates.)
Double-click an object to jump the 2D field to that point on the corresponding section.
Click meshes to select them. With objects selected: arrow keys translate in X/Y, Ctrl+Up/Down translates in Z, Shift+arrows and Ctrl+Shift+Up/Down rotate. Delete/Backspace removes selected objects from the scene; [ / ] step their opacity.
Set the camera along an axis with X / Y / Z, fit everything with Home ("Focus on all"), and center on the selection with F.
Press ? in the scene for its full shortcut list.

The 3D window has its own menu bar:

Scale cube: toggle a reference cube with the C key (or Scale Cube ▸ Display in scene). To move it, select it (left-click) and use the arrow keys (X/Y) and Ctrl+Up/Down (Z). Edit its edge length (µm), color, opacity, and outline width via Edit ▸ Edit attributes… (Ctrl+E). It's a physical measuring reference.
Scene ▸ Change background sets the background color.
Scene ▸ Organize scene… (Ctrl+Shift+H) lines objects up side by side, by host group or individually.
Scene ▸ Set translate/rotate step… sets the movement increments (default 0.1 µm and 10°).
File ▸ Save scene… / Load scene… save the scene (objects, colors, camera) to a JSON file and reload it later (also Series ▸ 3D ▸ Load 3D scene… from the main window). Add to scene can also pull objects from another series.

Saving images and exporting meshes¶

Scene ▸ Save scene screenshot… saves a rendered image (PNG, JPG, TIF, or BMP) at the configured DPI.
Scene ▸ Export scene… exports the whole scene as a single Wavefront .obj (with a .mtl material file).
From the Object list ▸ 3D ▸ Export mesh as, export individual objects as Wavefront (.obj), OFF (.off), Stanford PLY (.ply), STL (.stl), or Collada (.dae). (Surface and Spheres objects export; contour and tube types do not. Collada export additionally requires the optional pycollada package; if it is missing you'll be told, rather than seeing an error.)
Object list ▸ 3D ▸ Export quantitative data writes per-object surface area and volume to CSV. (Note these depend on the meshing settings; verify mesh quality before relying on the numbers.)

12. Saving and backups¶

Saving¶

File ▸ Save (Ctrl+S) writes the series back to its .jser. File ▸ Save as… writes a copy to a new location. PyReconstruct does not auto-save on a timer, so save regularly. (If you close with unsaved changes, you are prompted to save. If a session ends uncleanly, reopening offers to recover it; see Opening and creating a series.)

Backups¶

A backup is a complete copy of the series saved as an ordinary .jser into a folder you choose, with a filename you configure. There are two ways to make one:

File ▸ Backup ▸ Backup now… (Ctrl+Shift+B) saves the current data and writes a backup, optionally with a comment appended to the filename.
Automatic backups: enable Auto-backup (create backup on every save) in the backup settings, and a backup is written every time you Save or Save As.

File ▸ Backup ▸ Settings… configures the Backup Folder and the filename template. The template is assembled from optional parts joined by a delimiter: a prefix, the series code, the series name, the date and time (with customizable strftime patterns), the username, and a suffix. A live preview shows the resulting filename, and name collisions get a numeric suffix (-01, -02, …). The folder and the auto-backup toggle are remembered per series; the naming template is shared across all series. (The backup folder and naming are also available under Series ▸ Options ▸ Backup.)

When backups go in a subfolder named by the series code (the checkbox in File ▸ Backup ▸ Settings…, or {series} in the folder), codes that differ only by /, \ or _ share one subfolder, because each of those becomes _ in the folder name. On a case-insensitive filesystem, codes that differ only in upper and lower case also share one subfolder. Each backup file still gets its own name, so none is overwritten.

The Backup Settings dialog: This Series, My Defaults (all series) with the backup folder and auto-backup, and Backup Naming with the example backup file name.

Because backups are just .jser files, restore a backup by opening it with File ▸ Open series…. If the configured backup folder is missing at save time (for example a disconnected network drive), PyReconstruct prompts you to set it, or disables auto-backup until you do.

Tip. Keep your backup folder on separate storage from your working .jser, and turn on auto-backup for active projects.

13. Keyboard shortcuts¶

These are the defaults, in the groups and words of Help ▸ Shortcuts list (?), where you can rebind them. A key with Cmd or Option is written for macOS first; on Windows and Linux it is Ctrl or Alt, shown after it.

The Shortcuts list: each action with its shortcut in a box that can be edited or cleared, and Reset Defaults, Cancel and OK buttons.

General¶

Key	Action
`Delete/Backspace`	Delete selected traces. Remove last point when polyline tracing or using scissors. Delete selected entry in lists.
`? (Shift+/)`	Display keyboard shortcuts
`Cmd+,` (`Ctrl+,` on Windows/Linux)	View all options
`Page Up`	Display the next (higher) section
`Page Down`	Display the preceding (lower) section
`/`	Switch between current and last viewed section

View¶

Key	Action
`X`	Focus mode
`Cmd` (`Ctrl` on Windows/Linux)	Hold and click in focus mode to split or incorporate a trace
`H`	Hide trace layer
`A`	Show all traces (ignore hidden)
`I`	Hide image
`-`	Decrease brightness
`=`	Increase brightness
`[`	Decrease contrast
`]`	Increase contrast
`Space`	Section blend
`Home`	Set view to image

Field Interactions¶

Key	Action
`Cmd+A` (`Ctrl+A` on Windows/Linux)	Select all traces on section, or all rows in a focused data list
`Cmd+D` (`Ctrl+D` on Windows/Linux)	Deselect all traces on section, or all rows in a focused data list
`Cmd+Shift+I` (`Ctrl+Shift+I` on Windows/Linux)	Invert which traces are selected on section, or which rows in a focused data list
`Cmd+E` (`Ctrl+E` on Windows/Linux)	Edit attributes of selected trace(s)
`Cmd+Shift+R` (`Ctrl+Shift+R` on Windows/Linux)	Smooth selection (traces, or objects over the object list)
`Shift+M`	Merge selected traces
`Cmd+Shift+M` (`Ctrl+Shift+M` on Windows/Linux)	Merge attributes of selected traces
`Shift+H`	Hide selected traces
`Cmd+U` (`Ctrl+U` on Windows/Linux)	Unhide all hidden traces on current section
`Shift+G`	Modify current palette button to match attributes of first selected trace
`Cmd+Shift+G` (`Ctrl+Shift+G` on Windows/Linux)	Modify current palette button to match attributes and shape of first selected trace
`Cmd+Shift+U` (`Ctrl+Shift+U` on Windows/Linux)	Unlock current section
`Cmd+T` (`Ctrl+T` on Windows/Linux)	Modify transform on current section
`Cmd+\` (`Ctrl+\` on Windows/Linux)	Align by correlation
`Cmd+Shift+H` (`Ctrl+Shift+H` on Windows/Linux)	Set host(s) for selected trace(s)

Edit¶

Key	Action
`Cmd+Z` (`Ctrl+Z` on Windows/Linux)	Undo
`Cmd+Y` (`Ctrl+Y` on Windows/Linux)	Redo
`Cmd+C` (`Ctrl+C` on Windows/Linux)	Copy selected traces to clipboard
`Cmd+Option+C` (`Ctrl+Alt+C` on Windows/Linux)	Copy selected traces onto other sections
`Cmd+Option+X` (`Ctrl+Alt+X` on Windows/Linux)	Delete selected trace name(s) from chosen sections
`Cmd+Shift+D` (`Ctrl+Shift+D` on Windows/Linux)	Add selected object(s) to the 3D scene
`Cmd+X` (`Ctrl+X` on Windows/Linux)	Cut selected traces to clipboard
`Cmd+V` (`Ctrl+V` on Windows/Linux)	Paste clipboard traces into section
`Cmd+Shift+V` (`Ctrl+Shift+V` on Windows/Linux)	Apply attributes of copied traces to selected trace(s)

Navigate¶

Key	Action
`Cmd+F` (`Ctrl+F` on Windows/Linux)	Find first instance of a trace in the series
`Shift+F`	Find a trace on the current section
`Cmd+G` (`Ctrl+G` on Windows/Linux)	Go to a specific section number
`Cmd+K` (`Ctrl+K` on Windows/Linux)	Search the menus for a command

File¶

Key	Action
`Cmd+O` (`Ctrl+O` on Windows/Linux)	Open a series file
`Cmd+S` (`Ctrl+S` on Windows/Linux)	Save
`Cmd+Shift+B` (`Ctrl+Shift+B` on Windows/Linux)	Backup series
`Cmd+N` (`Ctrl+N` on Windows/Linux)	New series
`Cmd+Option+R` (`Ctrl+Alt+R` on Windows/Linux)	Restart
`Cmd+Q` (`Ctrl+Q` on Windows/Linux)	Quit

Lists¶

Key	Action
`Cmd+Option+S` (`Ctrl+Alt+S` on Windows/Linux)	Show or hide the lists
`Cmd+Shift+O` (`Ctrl+Shift+O` on Windows/Linux)	Open Object List
`Cmd+Shift+C` (`Ctrl+Shift+C` on Windows/Linux)	Toggle curation columns in object list
`Cmd+Shift+T` (`Ctrl+Shift+T` on Windows/Linux)	Open Trace List
`Cmd+Option+Z` (`Ctrl+Alt+Z` on Windows/Linux)	Open Z-trace List
`Cmd+Shift+S` (`Ctrl+Shift+S` on Windows/Linux)	Open Section List
`Cmd+Shift+F` (`Ctrl+Shift+F` on Windows/Linux)	Open Flag List
`Cmd+Shift+A` (`Ctrl+Shift+A` on Windows/Linux)	Switch/modify alignments

Trace Palette¶

Key	Action
`#, Shift+#`	Select a trace on the palette
`Cmd+#, Cmd+Shift+#` (`Ctrl+#, Ctrl+Shift+#` on Windows/Linux)	Edit attributes of a single trace on the palette
`Cmd+Shift+P` (`Ctrl+Shift+P` on Windows/Linux)	Switch/modify palettes
`Cmd+]` (`Ctrl+]` on Windows/Linux)	Increment palette {#} up
`Cmd+[` (`Ctrl+[` on Windows/Linux)	Increment palette {#} down

Movements¶

Key	Action
`Left/Right/Up/Down`	Translate selected traces or image (when no trace selected)
`Cmd+Left/Right/Up/Down` (`Ctrl+Left/Right/Up/Down` on Windows/Linux)	Translate traces or image by small step
`Shift+Left/Right/Up/Down`	Translate trace or image by large step
`Cmd+Shift+Left/Right/Up/Down` (`Ctrl+Shift+Left/Right/Up/Down` on Windows/Linux)	Rotate image around the mouse
`F1, Shift+F1, F2, Shift+F2`	Scale image in X and Y
`F3, Shift+F3, F4, Shift+F4`	Shear image in X and Y

Tool Palette¶

Key	Action
`P`	Use pointer tool
`Z`	Use pan/zoom tool
`K`	Use knife tool
`Shift+K`	Use scissors tool
`C`	Use closed trace tool
`O`	Use open trace tool
`S`	Use stamp tool
`G`	Use grid tool
`F`	Use flag tool
`Q`	Use host tool

3D Scene¶

Key	Action
`? (Shift+/)`	Display shortcuts in 3D scene

14. Getting help¶

In the app, Help ▸ Online resources links to:

the PyReconstruct user guide on the UT Austin lab wiki,
this distribution's source code,
the Kristen Harris Lab website and the Atlas of Ultrastructural Neurocytology.

Found a bug, have a feature idea, or want to improve the docs? Please open an issue on this distribution's GitHub Issues. The in-app Help ▸ Report a bug... and Help ▸ Request a feature... items open the same place with your version and OS already filled in. If you file another way, Help ▸ Copy diagnostic report gathers that in one paste, or click the version/commit shown at the top of the Help menu to copy the commit to your clipboard. Help ▸ Email developers writes to the same team.

Credits¶

PyReconstruct was created by Michael A. Chirillo, Julian N. Falco, Michael D. Musslewhite, Larry F. Lindsey, and Kristen M. Harris (Kristen Harris Lab, Department of Neuroscience, Center for Learning and Memory, The University of Texas at Austin) and introduced in PNAS (2025). The upstream project lives at SynapseWeb/PyReconstruct; it succeeds the original Reconstruct by John C. Fiala.

This distribution is independently developed and maintained by Dusten Hubbard (Kristen Harris Lab, The University of Texas at Austin). See the README for provenance, performance notes, and citation details, and CONTRIBUTING.md to help improve PyReconstruct.