vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

ChatON:ChatTemplateApplyCAPI remaining base logic

Browse source 
As c doesnt have the concept of pass by reference, and inturn the
existing c api uses pointers wrt llama chat message structure, so
switching to same wrt chat_tmpl_apply logics.

Also fix a oversight in previous commit and add the remaining logic.
This commit is contained in:
HanishKVC 2024-04-26 09:09:41 +05:30
parent 308d3bf3ff
commit a630564c48
1 changed files with 12 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

17
common/chaton.hpp
View file

@ -300,7 +300,7 @@ inline size_t chat_tmpl_apply_single_capi(
// NOTE: returns types and lens to help identify the parts of the tagged msg, which relate to passed and added tags // NOTE: returns types and lens to help identify the parts of the tagged msg, which relate to passed and added tags
inline bool chaton_tmpl_apply_ex( inline bool chaton_tmpl_apply_ex(
const std::string &tmpl, const std::string &tmpl,
const std::vector<llama_chat_message> &msgs, const std::vector<const llama_chat_message *> &msgs,
std::string &tagged, std::string &tagged,
std::string &types, std::string &types,
std::vector<int> &lens, std::vector<int> &lens,
@ -318,8 +318,8 @@ inline bool chaton_tmpl_apply_ex(
int cntUser = 0; int cntUser = 0;
int cntOthers = 0; int cntOthers = 0;
for(const auto msg: msgs) { for(const auto msg: msgs) {
auto role = msg.role; auto role = msg->role;
auto content = msg.content; auto content = msg->content;
std::string begin = chaton_tmpl_role_kv(tmpl, role, {K_BEGIN}); std::string begin = chaton_tmpl_role_kv(tmpl, role, {K_BEGIN});
auto prefix = chaton_tmpl_role_kv(tmpl, role, {K_PREFIX}); auto prefix = chaton_tmpl_role_kv(tmpl, role, {K_PREFIX});
auto suffix = chaton_tmpl_role_kv(tmpl, role, {K_SUFFIX}); auto suffix = chaton_tmpl_role_kv(tmpl, role, {K_SUFFIX});
@ -395,7 +395,7 @@ inline bool chaton_tmpl_apply_ex(
// then 1st user message will have user-prefix only if systemuser-1st-user-has-prefix is true // then 1st user message will have user-prefix only if systemuser-1st-user-has-prefix is true
inline int32_t chaton_tmpl_apply( inline int32_t chaton_tmpl_apply(
const std::string &tmpl, const std::string &tmpl,
const std::vector<llama_chat_message> &msgs, const std::vector<const llama_chat_message *> &msgs,
bool alertAssistantAtEnd, bool alertAssistantAtEnd,
std::string &tagged std::string &tagged
) { ) {
@ -420,10 +420,17 @@ inline int32_t chaton_tmpl_apply_capi(
} }
std::vector<const llama_chat_message *> vMsgs; std::vector<const llama_chat_message *> vMsgs;
for(size_t i=0; i<numMsgs; i++) { for(size_t i=0; i<numMsgs; i++) {
vMsgs.push_back(vMsgs[i]); vMsgs.push_back(&msgs[i]);
} }
std::string taggedMsgs; std::string taggedMsgs;
int32_t taggedLength = chaton_tmpl_apply(tmpl, vMsgs, alertAssistantAtEnd, taggedMsgs); int32_t taggedLength = chaton_tmpl_apply(tmpl, vMsgs, alertAssistantAtEnd, taggedMsgs);
if (taggedLength <= 0) {
return taggedLength;
}
if (destLength > 0) {
strlcpy(dest, taggedMsgs.c_str(), destLength);
}
return taggedLength;
} }