spec: Switch from llama-cli to llama-completion and llama-server

llama-cli as a different show off program now. So, switch to older
llama-completion for testing and to llama-server for versioning and
man-page generation.

Signed-off-by: Vitaly Chikunov <vt@altlinux.org>
This commit is contained in:
Vitaly Chikunov 2025-12-14 06:01:28 +03:00
parent 3ac9824627
commit e14c061238

View file

@ -176,7 +176,7 @@ export NVCC_PREPEND_FLAGS=-ccbin=g++-12
grep -E 'LLAMA|GGML' %_cmake__builddir/CMakeCache.txt | sort | tee build-options.txt
%cmake_build
find -name '*.py' | xargs sed -i '1s|#!/usr/bin/env python3|#!%__python3|'
LD_LIBRARY_PATH=%_cmake__builddir/bin %_cmake__builddir/bin/llama-cli --completion-bash > llama.bash
LD_LIBRARY_PATH=%_cmake__builddir/bin %_cmake__builddir/bin/llama-server --completion-bash > llama.bash
LD_LIBRARY_PATH=%_cmake__builddir/bin .gear/gen-manpage %_cmake__builddir/bin/llama-server > llama-server.1
%install
@ -192,9 +192,9 @@ install -Dp examples/*.py -t %buildroot%_datadir/%name/examples
# We need to run the tests, not install them.
rm %buildroot%_bindir/test-*
# Completions.
install -Dpm644 llama.bash %buildroot%_datadir/bash-completion/completions/llama-cli
printf '%%s\n' llama-server llama-simple llama-run llama-mtmd-cli |
xargs -ti ln -s llama-cli %buildroot%_datadir/bash-completion/completions/{}
install -Dpm644 llama.bash %buildroot%_datadir/bash-completion/completions/llama-server
printf '%%s\n' llama-cli llama-simple llama-run llama-mtmd-cli |
xargs -ti ln -s llama-server %buildroot%_datadir/bash-completion/completions/{}
mv %buildroot%_bindir/rpc-server %buildroot%_bindir/llama-rpc-server
install -Dpm644 llama-server.1 -t %buildroot%_man1dir
@ -204,12 +204,12 @@ install -Dpm644 llama-server.1 -t %buildroot%_man1dir
# Local path are more useful for debugging becasue they are not stripped by default.
%dnl export LD_LIBRARY_PATH=%buildroot%_libdir:%buildroot%_libexecdir/llama PATH+=:%buildroot%_bindir
export LD_LIBRARY_PATH=$PWD/%_cmake__builddir/bin PATH+=:$PWD/%_cmake__builddir/bin
llama-cli --version
llama-cli --version |& grep -Ex 'version: %version \(\S+ \[%release\]\)'
llama-server --version
llama-server --version |& grep -Ex 'version: %version \(\S+ \[%release\]\)'
# test-eval-callback wants network.
%ctest -j1 -E test-eval-callback
llama-cli -m %_datadir/tinyllamas/stories260K.gguf -p "Hello" -s 42 -n 500
llama-cli -m %_datadir/tinyllamas/stories260K.gguf -p "Once upon a time" -s 55 -n 33 |
llama-completion -m /usr/share/tinyllamas/stories260K.gguf -p "Hello" -s 42 -n 500 2>/dev/null
llama-completion -m /usr/share/tinyllamas/stories260K.gguf -p "Once upon a time" -s 55 -n 33 2>/dev/null |
grep 'Once upon a time, there was a boy named Tom. Tom had a big box of colors.'
# We do not provide convert tools.
mv %buildroot%_bindir/convert*.py -t %buildroot%_datadir/%name/examples