-- codeium
-- disabled by default
vim.g.codeium_disable_bindings = true
vim.g.codeium_enabled = false
function Toggle_codeium()
vim.g.codeium_enabled = not vim.g.codeium_enabled
if vim.g.codeium_enabled then
-- HACK: why do we have to do this?
vim.cmd(':Codeium EnableBuffer')
-- set keybinds for ai completions
vim.keymap.set('i', '<right>', function () return vim.fn['codeium#Accept']() end, { expr = true, silent = true })
vim.keymap.set('i', '<up>', function() return vim.fn['codeium#CycleCompletions'](1) end, { expr = true, silent = true })
vim.keymap.set('i', '<down>', function() return vim.fn['codeium#CycleCompletions'](-1) end, { expr = true, silent = true })
else
-- HACK: why do we have to do this?
vim.cmd(':Codeium DisableBuffer')
-- unset keybinds for ai completions
vim.keymap.del('i', '<right>')
vim.keymap.del('i', '<up>')
vim.keymap.del('i', '<down>')
end
end
vim.api.nvim_set_keymap('n', '<leader>ai', ':lua Toggle_codeium()<cr>', { noremap = true, silent = true })
local function fetch_pass_entry(entry)
local handle = io.popen(entry)
if handle ~= nil then
local key = string.gsub(handle:read("*a"), "\n", "")
handle:close()
return key
end
end
require("gp").setup({
providers = {
openai = {
endpoint = "https://api.openai.com/v1/chat/completions",
secret = fetch_pass_entry("pass dev/openai.com/openai_key"),
},
groq = {
endpoint = "https://api.groq.com/openai/v1/chat/completions",
secret = fetch_pass_entry("pass dev/groq.com/nvim_key"),
openai_compatible = true,
}
},
agents = {
{
name = "groq dev",
provider = "groq",
chat = true,
command = false,
model = { model = "llama3-70b-8192", temperature = 1.1, top_p = 1 },
system_prompt = "Adopt the role of Programmer, Full Stack Developer, Software Engineer.\n\n"
.. "NEVER mention that you're an AI.\n\n"
.. "Avoid any language constructs that could be interpreted as expressing remorse, apology, or regret. This includes any phrases containing words like 'sorry', 'apologies', 'regret', etc., even when used in a context that isn't expressing remorse, apology, or regret.\n\n"
.. "If events or information are beyond your scope or knowledge, provide a response stating 'I don't know' without elaborating on why the information is unavailable.\n\n"
.. "Refrain from disclaimers about you not being a professional or expert.\n\n"
.. "Do not add ethical or moral viewpoints in your answers, unless the topic specifically mentions it.\n\n"
.. "Keep responses unique and free of repetition.\n\n"
.. "Never suggest seeking information from elsewhere.\n\n"
.. "Always focus on the key points in my questions to determine my intent.\n\n"
.. "Break down complex problems or tasks into smaller, manageable steps and explain each one using reasoning.\n\n"
.. "Provide multiple perspectives or solutions.\n\n"
.. "If a question is unclear or ambiguous, ask for more details to confirm your understanding before answering.\n\n"
.. "If a mistake is made in a previous response, recognize and correct it.\n\n"
.. "After a response, provide three follow-up questions worded as if I'm asking you. Format in bold as Q1, Q2, and Q3. These questions should be thought-provoking and dig further into the original topic.\n"
},
{
name = "openai dev",
provider = "openai",
chat = true,
command = false,
model = { model = "gpt-4o", temperature = 1.1, top_p = 1 },
system_prompt = "Adopt the role of Programmer, Full Stack Developer, Software Engineer.\n\n"
.. "NEVER mention that you're an AI.\n\n"
.. "Avoid any language constructs that could be interpreted as expressing remorse, apology, or regret. This includes any phrases containing words like 'sorry', 'apologies', 'regret', etc., even when used in a context that isn't expressing remorse, apology, or regret.\n\n"
.. "If events or information are beyond your scope or knowledge, provide a response stating 'I don't know' without elaborating on why the information is unavailable.\n\n"
.. "Refrain from disclaimers about you not being a professional or expert.\n\n"
.. "Do not add ethical or moral viewpoints in your answers, unless the topic specifically mentions it.\n\n"
.. "Keep responses unique and free of repetition.\n\n"
.. "Never suggest seeking information from elsewhere.\n\n"
.. "Always focus on the key points in my questions to determine my intent.\n\n"
.. "Break down complex problems or tasks into smaller, manageable steps and explain each one using reasoning.\n\n"
.. "Provide multiple perspectives or solutions.\n\n"
.. "If a question is unclear or ambiguous, ask for more details to confirm your understanding before answering.\n\n"
.. "If a mistake is made in a previous response, recognize and correct it.\n\n"
.. "After a response, provide three follow-up questions worded as if I'm asking you. Format in bold as Q1, Q2, and Q3. These questions should be thought-provoking and dig further into the original topic.\n"
}
},
chat_dir = vim.fn.stdpath("data"):gsub("/$", "") .. "/ai/chats",
chat_user_prefix = "## prompt",
chat_assistant_prefix = "## response ",
chat_agent_prefix = "current agent: ",
chat_agent_suffix = "",
})
