mirror of
https://github.com/trustgraph-ai/trustgraph.git
synced 2026-07-17 09:11:03 +02:00
LLM migration to env secrets complete
This commit is contained in:
Cyber MacGeddon
parent
aab6472e3c
commit
0eee19b540
3 changed files with 55 additions and 84 deletions
|
|
@ -9,13 +9,15 @@ local prompts = import "prompts/mixtral.jsonnet";
|
|||
"chunk-size":: 150,
|
||||
"chunk-overlap":: 10,
|
||||
|
||||
"cohere-key":: "${COHERE_KEY}",
|
||||
"cohere-temperature":: 0.0,
|
||||
|
||||
"text-completion" +: {
|
||||
|
||||
create:: function(engine)
|
||||
|
||||
local envSecrets = engine.envSecrets("cohere-credentials")
|
||||
.with_env_var("COHERE_KEY", "cohere-key");
|
||||
|
||||
local container =
|
||||
engine.container("text-completion")
|
||||
.with_image(images.trustgraph)
|
||||
|
|
@ -23,42 +25,19 @@ local prompts = import "prompts/mixtral.jsonnet";
|
|||
"text-completion-cohere",
|
||||
"-p",
|
||||
url.pulsar,
|
||||
"-k",
|
||||
$["cohere-key"],
|
||||
"-t",
|
||||
std.toString($["cohere-temperature"]),
|
||||
])
|
||||
.with_limits("0.5", "128M")
|
||||
.with_reservations("0.1", "128M");
|
||||
|
||||
local containerSet = engine.containers(
|
||||
"text-completion", [ container ]
|
||||
);
|
||||
|
||||
local service =
|
||||
engine.internalService(containerSet)
|
||||
.with_port(8000, 8000, "metrics");
|
||||
|
||||
engine.resources([
|
||||
containerSet,
|
||||
service,
|
||||
])
|
||||
|
||||
},
|
||||
|
||||
"text-completion-rag" +: {
|
||||
|
||||
create:: function(engine)
|
||||
|
||||
local container =
|
||||
local containerRag =
|
||||
engine.container("text-completion-rag")
|
||||
.with_image(images.trustgraph)
|
||||
.with_command([
|
||||
"text-completion-cohere",
|
||||
"-p",
|
||||
url.pulsar,
|
||||
"-k",
|
||||
$["cohere-key"],
|
||||
"-t",
|
||||
std.toString($["cohere-temperature"]),
|
||||
"-i",
|
||||
|
|
@ -70,20 +49,30 @@ local prompts = import "prompts/mixtral.jsonnet";
|
|||
.with_reservations("0.1", "128M");
|
||||
|
||||
local containerSet = engine.containers(
|
||||
"text-completion-rag", [ container ]
|
||||
"text-completion", [ container ]
|
||||
);
|
||||
|
||||
local containerSetRag = engine.containers(
|
||||
"text-completion-rag", [ containerRag ]
|
||||
);
|
||||
|
||||
local service =
|
||||
engine.internalService(containerSet)
|
||||
.with_port(8000, 8000, "metrics");
|
||||
|
||||
local serviceRag =
|
||||
engine.internalService(containerSetRag)
|
||||
.with_port(8000, 8000, "metrics");
|
||||
|
||||
engine.resources([
|
||||
envSecrets,
|
||||
containerSet,
|
||||
containerSetRag,
|
||||
service,
|
||||
serviceRag,
|
||||
])
|
||||
|
||||
|
||||
}
|
||||
},
|
||||
|
||||
} + prompts
|
||||
|
||||
|
|
|
|||
|
|
@ -6,12 +6,14 @@ local prompts = import "prompts/slm.jsonnet";
|
|||
{
|
||||
|
||||
"ollama-model":: "gemma2:9b",
|
||||
"ollama-url":: "${OLLAMA_HOST}",
|
||||
|
||||
"text-completion" +: {
|
||||
|
||||
create:: function(engine)
|
||||
|
||||
local envSecrets = engine.envSecrets("ollama-credentials")
|
||||
.with_env_var("OLLAMA_HOST", "ollama-host");
|
||||
|
||||
local container =
|
||||
engine.container("text-completion")
|
||||
.with_image(images.trustgraph)
|
||||
|
|
@ -21,32 +23,12 @@ local prompts = import "prompts/slm.jsonnet";
|
|||
url.pulsar,
|
||||
"-m",
|
||||
$["ollama-model"],
|
||||
"-r",
|
||||
$["ollama-url"],
|
||||
])
|
||||
.with_env_var_secrets(envSecrets)
|
||||
.with_limits("0.5", "128M")
|
||||
.with_reservations("0.1", "128M");
|
||||
|
||||
local containerSet = engine.containers(
|
||||
"text-completion", [ container ]
|
||||
);
|
||||
|
||||
local service =
|
||||
engine.internalService(containerSet)
|
||||
.with_port(8080, 8080, "metrics");
|
||||
|
||||
engine.resources([
|
||||
containerSet,
|
||||
service,
|
||||
])
|
||||
|
||||
},
|
||||
|
||||
"text-completion-rag" +: {
|
||||
|
||||
create:: function(engine)
|
||||
|
||||
local container =
|
||||
local containerRag =
|
||||
engine.container("text-completion-rag")
|
||||
.with_image(images.trustgraph)
|
||||
.with_command([
|
||||
|
|
@ -55,31 +37,40 @@ local prompts = import "prompts/slm.jsonnet";
|
|||
url.pulsar,
|
||||
"-m",
|
||||
$["ollama-model"],
|
||||
"-r",
|
||||
$["ollama-url"],
|
||||
"-i",
|
||||
"non-persistent://tg/request/text-completion-rag",
|
||||
"-o",
|
||||
"non-persistent://tg/response/text-completion-rag-response",
|
||||
])
|
||||
.with_env_var_secrets(envSecrets)
|
||||
.with_limits("0.5", "128M")
|
||||
.with_reservations("0.1", "128M");
|
||||
|
||||
local containerSet = engine.containers(
|
||||
"text-completion-rag", [ container ]
|
||||
"text-completion", [ container ]
|
||||
);
|
||||
|
||||
local containerSetRag = engine.containers(
|
||||
"text-completion-rag", [ containerRag ]
|
||||
);
|
||||
|
||||
local service =
|
||||
engine.internalService(containerSet)
|
||||
.with_port(8080, 8080, "metrics");
|
||||
|
||||
local serviceRag =
|
||||
engine.internalService(containerSetRag)
|
||||
.with_port(8080, 8080, "metrics");
|
||||
|
||||
engine.resources([
|
||||
envSecrets,
|
||||
containerSet,
|
||||
containerSetRag,
|
||||
service,
|
||||
serviceRag,
|
||||
])
|
||||
|
||||
|
||||
}
|
||||
},
|
||||
|
||||
} + prompts
|
||||
|
||||
|
|
|
|||
|
|
@ -5,7 +5,6 @@ local prompts = import "prompts/mixtral.jsonnet";
|
|||
|
||||
{
|
||||
|
||||
"openai-key":: "${OPENAI_KEY}",
|
||||
"openai-max-output-tokens":: 4096,
|
||||
"openai-temperature":: 0.0,
|
||||
"openai-model":: "GPT-3.5-Turbo",
|
||||
|
|
@ -14,6 +13,9 @@ local prompts = import "prompts/mixtral.jsonnet";
|
|||
|
||||
create:: function(engine)
|
||||
|
||||
local envSecrets = engine.envSecrets("openai-credentials")
|
||||
.with_env_var("OPENAI_TOKEN", "openai-token");
|
||||
|
||||
local container =
|
||||
engine.container("text-completion")
|
||||
.with_image(images.trustgraph)
|
||||
|
|
@ -21,8 +23,6 @@ local prompts = import "prompts/mixtral.jsonnet";
|
|||
"text-completion-openai",
|
||||
"-p",
|
||||
url.pulsar,
|
||||
"-k",
|
||||
$["openai-key"],
|
||||
"-x",
|
||||
std.toString($["openai-max-output-tokens"]),
|
||||
"-t",
|
||||
|
|
@ -30,37 +30,17 @@ local prompts = import "prompts/mixtral.jsonnet";
|
|||
"-m",
|
||||
$["openai-model"],
|
||||
])
|
||||
.with_env_var_secrets(envSecrets)
|
||||
.with_limits("0.5", "128M")
|
||||
.with_reservations("0.1", "128M");
|
||||
|
||||
local containerSet = engine.containers(
|
||||
"text-completion", [ container ]
|
||||
);
|
||||
|
||||
local service =
|
||||
engine.internalService(containerSet)
|
||||
.with_port(8080, 8080, "metrics");
|
||||
|
||||
engine.resources([
|
||||
containerSet,
|
||||
service,
|
||||
])
|
||||
|
||||
},
|
||||
|
||||
"text-completion-rag" +: {
|
||||
|
||||
create:: function(engine)
|
||||
|
||||
local container =
|
||||
local containerRag =
|
||||
engine.container("text-completion-rag")
|
||||
.with_image(images.trustgraph)
|
||||
.with_command([
|
||||
"text-completion-openai",
|
||||
"-p",
|
||||
url.pulsar,
|
||||
"-k",
|
||||
$["openai-key"],
|
||||
"-x",
|
||||
std.toString($["openai-max-output-tokens"]),
|
||||
"-t",
|
||||
|
|
@ -72,24 +52,35 @@ local prompts = import "prompts/mixtral.jsonnet";
|
|||
"-o",
|
||||
"non-persistent://tg/response/text-completion-rag-response",
|
||||
])
|
||||
.with_env_var_secrets(envSecrets)
|
||||
.with_limits("0.5", "128M")
|
||||
.with_reservations("0.1", "128M");
|
||||
|
||||
local containerSet = engine.containers(
|
||||
"text-completion-rag", [ container ]
|
||||
"text-completion", [ container ]
|
||||
);
|
||||
|
||||
local containerSetRag = engine.containers(
|
||||
"text-completion-rag", [ containerRag ]
|
||||
);
|
||||
|
||||
local service =
|
||||
engine.internalService(containerSet)
|
||||
.with_port(8080, 8080, "metrics");
|
||||
|
||||
local serviceRag =
|
||||
engine.internalService(containerSetRag)
|
||||
.with_port(8080, 8080, "metrics");
|
||||
|
||||
engine.resources([
|
||||
envSecrets,
|
||||
containerSet,
|
||||
containerSetRag,
|
||||
service,
|
||||
serviceRag,
|
||||
])
|
||||
|
||||
|
||||
}
|
||||
},
|
||||
|
||||
} + prompts
|
||||
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue