macOS Bash Shell Integration¶

Default shell before macOS Catalina (10.15). Since Catalina, macOS ships with Zsh as the default shell. If you are on a modern Mac and have not changed your shell, see the macOS Zsh Integration guide instead.

You can check your current shell with:
echo $SHELL

To make interacting with the STT (Speech-to-Text) CLI easier, you can add a shortcut function to your ~/.bash_profile. This allows you to simply type s "your question" in the terminal.

Setup Instructions¶

Open your Bash configuration with an editor you like:

nano ~/.bash_profile
open -e ~/.bash_profile   # opens in TextEdit

Paste the following block at the end of the file:

please read newest updates in zsh - verson


# --- STT Project Path Resolution ---
unalias s 2>/dev/null
s() {
    if [ $# -eq 0 ]; then
        echo "question <your question>"
        return 1
    fi
    update_github_ip
    local TEMP_FILE=$(mktemp)
    local SHORT_TIMEOUT_SECONDS=2
    local LONG_TIMEOUT_SECONDS=70
    # Path shortcuts
    local PY_EXEC="$SL5NET_AURA_PROJECT_ROOT/.venv/bin/python3"
    local CLI_SCRIPT="$SL5NET_AURA_PROJECT_ROOT/scripts/py/cli_client.py"
    # --- 1. try
    timeout $SHORT_TIMEOUT_SECONDS \
    "$PY_EXEC" -u "$CLI_SCRIPT" "$*" \
    --lang "de-DE" --unmasked < /dev/null > "$TEMP_FILE" 2>&1
    local EXIT_CODE=$?
    local OUTPUT=$(cat "$TEMP_FILE")
    rm "$TEMP_FILE"
    if echo "$OUTPUT" | grep -q "Verbindungsfehler" || ! pgrep -f "streamlit-chat.py" > /dev/null; then
        echo "Service-Check: Backend oder Frontend fehlt. Starte neu..."
        start_service
        echo '++++++++++++++++++++++++++++++++++++++++++++++++++'
        local KIWIX_SCRIPT="$SL5NET_AURA_PROJECT_ROOT/config/maps/plugins/standard_actions/wikipedia_local/de-DE/kiwix-docker-start-if-not-running.sh"
        if [ -f "$KIWIX_SCRIPT" ]; then
            bash "$KIWIX_SCRIPT"
        fi
        echo '++++++++++++++++++++++++++++++++++++++++++++++++++'
        echo "BITTE ERNEUT EINGEBEN: s $*"
        return 1
    # 2. Timeout (124) OR success (0)
    elif [ $EXIT_CODE -eq 124 ] || [ $EXIT_CODE -eq 0 ]; then
        if [ $EXIT_CODE -eq 0 ]; then
            echo "$OUTPUT"
            return 0
        fi
        echo "answer > $SHORT_TIMEOUT_SECONDS sec. set Timeout= $LONG_TIMEOUT_SECONDS s..."
        local TEMP_FILE_2=$(mktemp)
        timeout $LONG_TIMEOUT_SECONDS \
        "$PY_EXEC" -u "$CLI_SCRIPT" "$*" \
        --lang "de-DE" --unmasked < /dev/null > "$TEMP_FILE_2" 2>&1
        local EXIT_CODE_2=$?
        local OUTPUT_2=$(cat "$TEMP_FILE_2")
        rm "$TEMP_FILE_2"
        echo "$OUTPUT_2"
        if [ $EXIT_CODE_2 -ne 0 ]; then
             echo "WARNUNG: Timeout > $LONG_TIMEOUT_SECONDS Sec. "
        fi
        return 0
    else
        echo "ERROR"
        echo "$OUTPUT"
        return $EXIT_CODE
    fi
}

Reload your configuration:
```
source ~/.bash_profile
```

macOS-Specific Notes¶

timeout is not built into macOS. Install it via Homebrew before using this function:
```
brew install coreutils
```
After installation, timeout is available as gtimeout. Either add an alias or replace timeout with gtimeout in the function above:
```
alias timeout=gtimeout
```
Add the alias above the s() function in your ~/.bash_profile.
macOS uses ~/.bash_profile for login shells (Terminal.app opens login shells by default), whereas Linux typically uses ~/.bashrc. If you want the function available in all contexts, you can source one from the other:
```
# Add to ~/.bash_profile:
[ -f ~/.bashrc ] && source ~/.bashrc
```
macOS ships with Bash 3.2 (due to the GPLv3 license). This function is fully compatible with Bash 3.2+. If you need Bash 5, install it via Homebrew:
```
brew install bash
```
Python path: Make sure your virtual environment is set up at $SL5NET_AURA_PROJECT_ROOT/.venv. If you manage Python with pyenv or conda, adjust PY_EXEC accordingly.

Features¶

Dynamic Paths: Automatically finds the project root via the /tmp marker file.
Auto-Restart: If the backend is down, it attempts to run start_service and local Wikipedia services.
Smart Timeouts: Tries a quick 2-second response first, then falls back to a 70-second deep processing mode.