Harvey Build Environment Setup

This document outlines the necessary steps to set up the build environment for the Harvey application on a Debian-based Linux system.

Important Note: These commands must be executed inside the specific build environment (e.g., within the Docker container, VM, or CI/CD runner) where the application build will take place.

1. System Dependencies

First, ensure the package list is up to date within your build environment:

sudo apt-get update

Next, install the required libraries and build tools for Tauri. The command below installs the core system dependencies:

sudo apt-get install -y \
    build-essential \
    curl \
    wget \
    file \
    libssl-dev \
    libglib2.0-dev \
    libgtk-3-dev \
    libayatana-appindicator3-dev \
    librsvg2-dev \
    libsoup-3.0-dev \
    libwebkit2gtk-4.1-dev

2. Python Virtual Environment

The application backend creates and manages a self-contained Python virtual environment (

harvey_env

) within the app's configuration directory. This environment is used for running tasks that rely on Python libraries, such as audio processing and document conversion.

Core dependencies are installed automatically by the application on first launch or when required. Key libraries include:

```
torch
```
and
```
torchaudio
```
for audio processing.
```
pyannote.audio
```
for speaker diarization.
```
pypandoc
```
for document conversion (or
```
pypandoc_binary
```
on some platforms).

No manual Python setup is required by the user for most platforms, as the application handles the creation and management of this environment using a bundled version of Micromamba.

Special Case: Windows ARM64

Due to the lack of Micromamba support for Windows on ARM64, the setup for this platform is different:

Python Environment: The build process (
```
build.rs
```
) automatically downloads and bundles a standalone version of Python from
```
astral-sh/python-build-standalone
```
. At runtime, the application creates a standard virtual environment from this bundled Python.
FFmpeg: This dependency is not bundled. The user must manually download and install FFmpeg from the official website and ensure that the
```
ffmpeg.exe
```
binary is available in their system's
```
PATH
```
. The application will detect if it's missing and prompt the user to install it.
Python Libraries: The list of Python libraries for Windows ARM64 is slightly different. Notably, it uses
```
pypandoc_binary
```
instead of
```
pypandoc
```
, and the
```
torch
```
libraries are installed without version constraints to pull the correct platform-specific wheels. This is all handled automatically by the application.

3. Building the Application

After installing the dependencies inside the build environment, you can build the Tauri application:

npm run tauri build

3. Troubleshooting

If the build fails with an error like "The system library ... was not found," it almost always means the dependency was not installed correctly inside the build environment.

Step A: Verify the Installation

Before trying anything else, verify that the required library is actually visible to the

pkg-config

tool from your build shell:

# This command should output a version number (e.g., 2.44.2).
# If it says the package was not found, the installation in step 1 failed or was done in the wrong environment.
pkg-config --modversion webkit2gtk-4.1

Step B: Check

PKG_CONFIG_PATH

On standard Debian systems, setting

PKG_CONFIG_PATH

is not usually necessary if the

-dev

packages were installed correctly. However, if the verification in Step A passed but the build still fails, you can explicitly set the path as a last resort:

export PKG_CONFIG_PATH=/usr/lib/x86_64-linux-gnu/pkgconfig
npm run tauri build

Harvey Build Environment Setup

This document outlines the necessary steps to set up the build environment for the Harvey application on a Debian-based Linux system.

Important Note: These commands must be executed inside the specific build environment (e.g., within the Docker container, VM, or CI/CD runner) where the application build will take place.

1. System Dependencies

First, ensure the package list is up to date within your build environment:

sudo apt-get update

Next, install the required libraries and build tools for Tauri. The command below installs the core system dependencies:

sudo apt-get install -y \
    build-essential \
    curl \
    wget \
    file \
    libssl-dev \
    libglib2.0-dev \
    libgtk-3-dev \
    libayatana-appindicator3-dev \
    librsvg2-dev \
    libsoup-3.0-dev \
    libwebkit2gtk-4.1-dev

2. Python Virtual Environment

The application backend creates and manages a self-contained Python virtual environment (

harvey_env

) within the app's configuration directory. This environment is used for running tasks that rely on Python libraries, such as audio processing and document conversion.

Core dependencies are installed automatically by the application on first launch or when required. Key libraries include:

```
torch
```
and
```
torchaudio
```
for audio processing.
```
pyannote.audio
```
for speaker diarization.
```
pypandoc
```
for document conversion (or
```
pypandoc_binary
```
on some platforms).

No manual Python setup is required by the user for most platforms, as the application handles the creation and management of this environment using a bundled version of Micromamba.

Special Case: Windows ARM64

Due to the lack of Micromamba support for Windows on ARM64, the setup for this platform is different:

Python Environment: The build process (
```
build.rs
```
) automatically downloads and bundles a standalone version of Python from
```
astral-sh/python-build-standalone
```
. At runtime, the application creates a standard virtual environment from this bundled Python.
FFmpeg: This dependency is not bundled. The user must manually download and install FFmpeg from the official website and ensure that the
```
ffmpeg.exe
```
binary is available in their system's
```
PATH
```
. The application will detect if it's missing and prompt the user to install it.
Python Libraries: The list of Python libraries for Windows ARM64 is slightly different. Notably, it uses
```
pypandoc_binary
```
instead of
```
pypandoc
```
, and the
```
torch
```
libraries are installed without version constraints to pull the correct platform-specific wheels. This is all handled automatically by the application.

3. Building the Application

After installing the dependencies inside the build environment, you can build the Tauri application:

npm run tauri build

3. Troubleshooting

If the build fails with an error like "The system library ... was not found," it almost always means the dependency was not installed correctly inside the build environment.

Step A: Verify the Installation

Before trying anything else, verify that the required library is actually visible to the

pkg-config

tool from your build shell:

# This command should output a version number (e.g., 2.44.2).
# If it says the package was not found, the installation in step 1 failed or was done in the wrong environment.
pkg-config --modversion webkit2gtk-4.1

Step B: Check

PKG_CONFIG_PATH

On standard Debian systems, setting

PKG_CONFIG_PATH

is not usually necessary if the

-dev

packages were installed correctly. However, if the verification in Step A passed but the build still fails, you can explicitly set the path as a last resort:

export PKG_CONFIG_PATH=/usr/lib/x86_64-linux-gnu/pkgconfig
npm run tauri build

Harvey Build Environment Setup

Harvey Build Environment Setup

1. System Dependencies

2. Python Virtual Environment

Special Case: Windows ARM64

3. Building the Application

3. Troubleshooting

Step A: Verify the Installation

Step B: Check
`PKG_CONFIG_PATH`

Related Skills

Nano Banana Pro

Markdown Converter

1password

Harvey Build Environment Setup

1. System Dependencies

2. Python Virtual Environment

Special Case: Windows ARM64

3. Building the Application

3. Troubleshooting

Step A: Verify the Installation

Step B: Check
`PKG_CONFIG_PATH`