Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly
Sign up
Appearance settings

Commit c48f630

Browse files
authored
llama : add --completion-bash option (ggml-org#11846)
This commit adds a new option `--completion-bash` to the llama.cpp which outputs a source-able bash completion script. The motivation for this change is to provide a more user-friendly experience for users who use the command-line interface of llama.cpp. This is currently only basic and all options are displayed for all llama executables but this can be improved in the future if needed. Example usage: ```console $ build/bin/llama-cli --completion-bash > ~/.llama-completion.bash $ source ~/.llama-completion.bash $ ./build/bin/llama-server --m<TAB> --main-gpu --mirostat --mirostat-lr --model --multiline-input --min-p --mirostat-ent --mlock --model-url ```
1 parent bd6e55b commit c48f630

File tree

3 files changed

+128
-0
lines changed

3 files changed

+128
-0
lines changed

‎README.md‎

Lines changed: 14 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -521,3 +521,17 @@ If your issue is with model generation quality, then please at least scan the fo
521521
522522
#### References
523523
524+
525+
### Completions
526+
Command-line completion is available for some environments.
527+
528+
#### Bash Completion
529+
```bash
530+
$ build/bin/llama-cli --completion-bash > ~/.llama-completion.bash
531+
$ source ~/.llama-completion.bash
532+
```
533+
Optionally this can be added to your `.bashrc` or `.bash_profile` to load it
534+
automatically. For example:
535+
```console
536+
$ echo "source ~/.llama-completion.bash" >> ~/.bashrc
537+
```

‎common/arg.cpp‎

Lines changed: 113 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -365,6 +365,108 @@ static void common_params_print_usage(common_params_context & ctx_arg) {
365365
print_options(specific_options);
366366
}
367367

368+
static void common_params_print_completion(common_params_context & ctx_arg) {
369+
std::vector<common_arg *> common_options;
370+
std::vector<common_arg *> sparam_options;
371+
std::vector<common_arg *> specific_options;
372+
373+
for (auto & opt : ctx_arg.options) {
374+
if (opt.is_sparam) {
375+
sparam_options.push_back(&opt);
376+
} else if (opt.in_example(ctx_arg.ex)) {
377+
specific_options.push_back(&opt);
378+
} else {
379+
common_options.push_back(&opt);
380+
}
381+
}
382+
383+
printf("_llama_completions() {\n");
384+
printf(" local cur prev opts\n");
385+
printf(" COMPREPLY=()\n");
386+
printf(" cur=\"${COMP_WORDS[COMP_CWORD]}\"\n");
387+
printf(" prev=\"${COMP_WORDS[COMP_CWORD-1]}\"\n\n");
388+
389+
printf(" opts=\"");
390+
auto print_options = [](const std::vector<common_arg *> & options) {
391+
for (const common_arg * opt : options) {
392+
for (const char * arg : opt->args) {
393+
printf("%s ", arg);
394+
}
395+
}
396+
};
397+
398+
print_options(common_options);
399+
print_options(sparam_options);
400+
print_options(specific_options);
401+
printf("\"\n\n");
402+
403+
printf(" case \"$prev\" in\n");
404+
printf(" --model)\n");
405+
printf(" COMPREPLY=( $(compgen -f -X '!*.gguf' -- \"$cur\") $(compgen -d -- \"$cur\") )\n");
406+
printf(" return 0\n");
407+
printf(" ;;\n");
408+
printf(" --grammar-file)\n");
409+
printf(" COMPREPLY=( $(compgen -f -X '!*.gbnf' -- \"$cur\") $(compgen -d -- \"$cur\") )\n");
410+
printf(" return 0\n");
411+
printf(" ;;\n");
412+
printf(" *)\n");
413+
printf(" COMPREPLY=( $(compgen -W \"${opts}\" -- \"$cur\") )\n");
414+
printf(" return 0\n");
415+
printf(" ;;\n");
416+
printf(" esac\n");
417+
printf("}\n\n");
418+
419+
std::set<std::string> executables = {
420+
"llama-batched",
421+
"llama-batched-bench",
422+
"llama-bench",
423+
"llama-cli",
424+
"llama-convert-llama2c-to-ggml",
425+
"llama-cvector-generator",
426+
"llama-embedding",
427+
"llama-eval-callback",
428+
"llama-export-lora",
429+
"llama-gbnf-validator",
430+
"llama-gen-docs",
431+
"llama-gguf",
432+
"llama-gguf-hash",
433+
"llama-gguf-split",
434+
"llama-gritlm",
435+
"llama-imatrix",
436+
"llama-infill",
437+
"llama-llava-cli",
438+
"llama-llava-clip-quantize-cli",
439+
"llama-lookahead",
440+
"llama-lookup",
441+
"llama-lookup-create",
442+
"llama-lookup-merge",
443+
"llama-lookup-stats",
444+
"llama-minicpmv-cli",
445+
"llama-parallel",
446+
"llama-passkey",
447+
"llama-perplexity",
448+
"llama-q8dot",
449+
"llama-quantize",
450+
"llama-quantize-stats",
451+
"llama-qwen2vl-cli",
452+
"llama-retrieval",
453+
"llama-run",
454+
"llama-save-load-state",
455+
"llama-server",
456+
"llama-simple",
457+
"llama-simple-chat",
458+
"llama-speculative",
459+
"llama-speculative-simple",
460+
"llama-tokenize",
461+
"llama-tts",
462+
"llama-vdot"
463+
};
464+
465+
for (const auto& exe : executables) {
466+
printf("complete -F _llama_completions %s\n", exe.c_str());
467+
}
468+
}
469+
368470
static std::vector<ggml_backend_dev_t> parse_device_list(const std::string & value) {
369471
std::vector<ggml_backend_dev_t> devices;
370472
auto dev_names = string_split<std::string>(value, ',');
@@ -426,6 +528,10 @@ bool common_params_parse(int argc, char ** argv, common_params & params, llama_e
426528
}
427529
exit(0);
428530
}
531+
if (ctx_arg.params.completion) {
532+
common_params_print_completion(ctx_arg);
533+
exit(0);
534+
}
429535
} catch (const std::invalid_argument & ex) {
430536
fprintf(stderr, "%s\n", ex.what());
431537
ctx_arg.params = params_org;
@@ -494,6 +600,13 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
494600
exit(0);
495601
}
496602
));
603+
add_opt(common_arg(
604+
{"--completion-bash"},
605+
"print source-able bash completion script for llama.cpp",
606+
[](common_params & params) {
607+
params.completion = true;
608+
}
609+
));
497610
add_opt(common_arg(
498611
{"--verbose-prompt"},
499612
string_format("print a verbose prompt before generation (default: %s)", params.verbose_prompt ? "true" : "false"),

‎common/common.h‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -298,6 +298,7 @@ struct common_params {
298298
bool kl_divergence = false; // compute KL divergence
299299

300300
bool usage = false; // print usage
301+
bool completion = false; // print source-able completion script
301302
bool use_color = false; // use color to distinguish generations and inputs
302303
bool special = false; // enable special token output
303304
bool interactive = false; // interactive mode

0 commit comments

Comments
(0)

AltStyle によって変換されたページ (->オリジナル) /