FusionExtractor

A Python library for extracting resources from Autodesk Fusion Electronics .f3z project files.

Overview

.f3z files are ZIP archives that embed multiple nested archives containing schematic (.sch), PCB (.brd), and preview image data. FusionExtractor unpacks these without any third-party dependencies.

Installation

pip install fusionextractor                  # stdlib only
pip install "fusionextractor[zstd]"          # adds Zstandard support for 3D model previews

Examples

A sample Fusion Electronics project is included in examples/IOMOD-AD5593R_v2_0.f3z for use by automatic tests and for you to use when trying out the library.

To run examples/extract.py, open a terminal in the library folder and type:

pip install -e ".[zstd]"
python examples/extract.py examples/IOMOD-AD5593R-v2_0.f3z

This will create a new folder called output in the current directory, and extract various assets from the sample project into it. Check the source of extract.py to see the options it uses relating to output filenames, etc.

Usage

from fusionextractor import FusionProject

with FusionProject("design.f3z") as proj:
    # Read the design name from embedded metadata
    print(proj.design_name)

    # Get raw bytes
    sch_bytes = proj.get_schematic()   # Eagle .sch file
    brd_bytes = proj.get_board()       # Eagle .brd file
    previews   = proj.get_previews()    # list[PreviewImage]
    bom        = proj.get_bom()         # list[BomEntry]
    placements = proj.get_placements()  # list[PlacementEntry]

    # Extract to disk
    proj.extract_schematic("output/")           # writes original filename into output/
    proj.extract_board("output/my_board.brd")   # writes to exact path
    proj.extract_previews("output/previews/")   # writes all preview PNGs
    proj.extract_bom("output/")                 # writes {design_name}_bom.csv
    proj.extract_placements_kicad("output/")    # writes {design_name}.pos
    proj.extract_placements_eagle("output/")    # writes {design_name}.mnt and {design_name}.mnb

    # Extract named board images
    proj.extract_board_image("schematic",    "output/")   # {design_name}_schematic.png
    proj.extract_board_image("pcb_top",      "output/")   # {design_name}_pcb_top.png
    proj.extract_board_image("pcb_3d_top",   "output/")   # {design_name}_pcb_3d_top.png
    proj.extract_board_image("pcb_3d_bottom","output/")   # {design_name}_pcb_3d_bottom.png

Destination path behaviour

extract_schematic and extract_board accept an optional dest argument:

dest value	Result
None (default)	Written to the current directory, original filename preserved
A directory path or path ending with /	Original filename preserved inside that directory
A full file path (with extension)	Written to that exact path

Board images

get_board_image and extract_board_image retrieve specific named views of the board:

view_type	Content
"schematic"	Schematic diagram (one or more pages)
"pcb_top"	PCB top-layer 2D layout
"pcb_3d_top"	3D render from the top
"pcb_3d_bottom"	3D render from the bottom

top_png    = proj.get_board_image("pcb_3d_top")    # returns bytes
bottom_png = proj.get_board_image("pcb_3d_bottom")

proj.extract_board_image("pcb_3d_top", "output/")  # writes {design_name}_pcb_3d_top.png
proj.extract_board_image("pcb_3d_top", "render.png")  # writes to exact path

Preview images

get_previews and extract_previews return all images from all nested archives, including small thumbnails and the large renders above. Each PreviewImage has a view_type field populated automatically.

Pass include_large_images=False to retrieve thumbnails only.

for preview in proj.get_previews(include_large_images=False):
    print(preview.source, len(preview.data), "bytes")
    # e.g. "schematic 4821 bytes"

PreviewImage fields:

Field	Type	Description
source	str	Archive the image came from ("schematic", "board", "project", "3d_model")
path	str	Path of the image inside the nested archive
data	bytes	Raw PNG bytes
view_type	str | None	"schematic", "pcb_top", "pcb_3d_top", "pcb_3d_bottom", or "thumbnail"

Extracted preview files are named {source}__{original_filename} to avoid collisions when multiple archives contain a small.png.

API reference

class FusionProject:
    path: str | Path                                          # path to the .f3z file

    design_name: str                                          # property; from embedded metadata

    def get_schematic() -> bytes
    def get_board() -> bytes
    def get_previews(*, include_large_images: bool = True) -> list[PreviewImage]
    def get_board_image(view_type: str) -> bytes              # "schematic"|"pcb_top"|"pcb_3d_top"|"pcb_3d_bottom"
    def get_bom(*, include_power_symbols: bool = False) -> list[BomEntry]
    def get_placements() -> list[PlacementEntry]

    def extract_schematic(dest=None) -> Path
    def extract_board(dest=None) -> Path
    def extract_previews(dest=None, *, include_large_images: bool = True) -> list[Path]
    def extract_board_image(view_type: str, dest=None) -> Path
    def extract_bom(dest=None, *, include_power_symbols: bool = False) -> Path
    def extract_placements_kicad(dest=None) -> Path
    def extract_placements_eagle(dest=None) -> tuple[Path, Path]

BOM

get_bom parses the component list from the embedded schematic. Supply/power symbols (GND, VCC, NC, etc.) are excluded by default; pass include_power_symbols=True to include them.

for entry in proj.get_bom():
    print(entry.reference, entry.device, entry.package, entry.value)
    # e.g. "C1  CAP  0603  100nF"

BomEntry fields:

Field	Type	Description
reference	str	Reference designator ("C1", "U1")
device	str	Component type / part number base ("CAP", "AD5593R")
package	str	Footprint / package ("0603", "TSSOP16")
value	str	Component value ("100nF", "10K") — empty string when not set
library	str	Eagle library name ("SuperHouse-Capacitors")

extract_bom writes the BOM as a CSV file named {design_name}_bom.csv.

Placements

get_placements parses component placement data (position, rotation, side) from the embedded board.

for entry in proj.get_placements():
    print(entry.reference, entry.x, entry.y, entry.rotation, entry.side)
    # e.g. "R1  6.0  4.0  90.0  top"

PlacementEntry fields:

Field	Type	Description
reference	str	Reference designator ("R1", "U1")
value	str	Component value ("100nF", "10K") — empty string when not set
package	str	Footprint ("C0603", "TSSOP16")
x	float	X position in mm, Eagle's native board coordinates (Y increasing upward)
y	float	Y position in mm, Eagle's native board coordinates (Y increasing upward)
rotation	float	Rotation in degrees (0-360)
side	str	"top" or "bottom"

Two extraction formats are available:

• extract_placements_kicad writes a single KiCad-format .pos file covering both sides. Because KiCad's PCB editor uses a Y-increases-downward axis while Eagle uses Y-increases-upward, PosY is the negation of PlacementEntry.y.
• extract_placements_eagle writes a pair of Eagle mount.ulp-format files split by side: {design_name}.mnt (top) and {design_name}.mnb (bottom). X/Y are written unchanged.

proj.extract_placements_kicad("output/")        # output/{design_name}.pos
mnt_path, mnb_path = proj.extract_placements_eagle("output/")  # .mnt + .mnb

Exceptions

Exception	Raised when
FileNotFoundError	The .f3z file path does not exist
FusionExtractorError	The file is not a valid ZIP/f3z archive
FileNotFoundInArchiveError	A required entry is missing from inside the archive

Both custom exceptions are subclasses of FusionExtractorError and are importable from the package root:

from fusionextractor import FusionExtractorError, FileNotFoundInArchiveError

Requirements

Python 3.9+. No required third-party dependencies.

The .f3d archive (3D model) uses Zstandard compression, which Python's stdlib zipfile does not support. Without the optional extra, 3D model previews are silently skipped. Install .[zstd] to enable them.