ollama bash completion script based on https://www.perplexity.ai/search/did-people-make-for-ollama-bas-3RErncU0QcqHwaS64ukerw

ollama_bash_completion_script.sh

#!/bin/bash
# Bash completion for the "ollama" CLI.  Requires jq and curl.
#
# Published: 
# * https://github.com/CLIAI/handy_scripts/blob/main/bash_completion/etc/bash_completion.d/ollama
# * https://gist.github.com/gwpl/a5e47034029bcb37de5b13f64108fc80/

# Cache settings
_OLLAMA_MODEL_TTL=300
_OLLAMA_MODELS_TIMESTAMP=0
_OLLAMA_MODELS=""

# Fetch models from Ollama server, caching results
_ollama_fetch_models() {
  local now
  now=$(date +%s)
  if [ $(( now - _OLLAMA_MODELS_TIMESTAMP )) -gt $_OLLAMA_MODEL_TTL ]; then
    _OLLAMA_MODELS=$(
      curl -s http://localhost:11434/api/tags \
      | jq -r '.models[].name'
    )
    _OLLAMA_MODELS_TIMESTAMP=$now
  fi
}

# Main completion function
_ollama() {
  local cur prev
  COMPREPLY=()
  cur="${COMP_WORDS[COMP_CWORD]}"
  prev="${COMP_WORDS[COMP_CWORD-1]}"

  case "${prev}" in
    run)
      # Get fresh models from Ollama
      _ollama_fetch_models

      # Filter models by whatever the user has typed so far (case-insensitive)
      # and feed them into compgen so they appear as completions.
      local filtered
      filtered=$(echo "$_OLLAMA_MODELS" | grep -i "$cur")
      COMPREPLY=( $(compgen -W "${filtered}" -- "$cur") )
      return 0
      ;;
  esac

  # Default commands
  COMPREPLY=( $(compgen -W "serve create show run pull push list ps cp rm help" -- "$cur") )
}

# Register the completion function
complete -F _ollama ollama

x_GenAI_Ollama_Bash_Completion_HowTo.md

Implementing Bash Autocomplete for Ollama

source: https://www.perplexity.ai/search/did-people-make-for-ollama-bas-3RErncU0QcqHwaS64ukerw

You can implement bash autocompletion for Ollama model names by creating a custom bash completion script. Based on the search results, there's already a well-developed solution available that provides exactly this functionality.

Installation Steps

1. Create a new file for the Ollama completion script:

sudo nano /etc/bash_completion.d/ollama

Or create it in your home directory:

nano ~/.ollama-completion.sh

2. Copy the completion script from the GitHub issue[1] into this file (script details below).

3. If you placed it in your home directory, add this line to your .bashrc file:

source ~/.ollama-completion.sh

4. Install required dependencies:

sudo apt install jq curl

5. Restart your terminal or source your bashrc file:

source ~/.bashrc

Trying to merge into Ollama mainstream repo: ollama/ollama#9585