Formatted src/data_generation/llama

tooling/pipeline/src/data_generation/llama/generate_brewery.cc

@@ -89,10 +89,10 @@ BreweryResult LlamaGenerator::GenerateBrewery(
       raw = this->Infer(system_prompt, user_prompt, max_tokens,
                         kBreweryJsonGrammar);
       if (logger_) {
-      logger_->Log(
+         logger_->Log({.level = LogLevel::Debug,
-          {.level = LogLevel::Debug,
                        .phase = PipelinePhase::BreweryAndBeerGeneration,
-           .message = std::format("LlamaGenerator: raw output (attempt {}): {}",
+                       .message = std::format(
+                           "LlamaGenerator: raw output (attempt {}): {}",
                            attempt + 1, raw)});
       }
 
@@ -106,10 +106,10 @@ BreweryResult LlamaGenerator::GenerateBrewery(
          // Success: return parsed brewery data
 
          if (logger_) {
-        logger_->Log(
+            logger_->Log({.level = LogLevel::Info,
-            {.level = LogLevel::Info,
                           .phase = PipelinePhase::BreweryAndBeerGeneration,
-             .message = std::format("LlamaGenerator: successfully generated "
+                          .message = std::format(
+                              "LlamaGenerator: successfully generated "
                               "brewery data on attempt {}",
                               attempt + 1)});
          }

tooling/pipeline/src/data_generation/llama/generate_user.cc

@@ -42,8 +42,8 @@ UserResult LlamaGenerator::GenerateUser(const EnrichedCity& city,
        style_affinities);
 
    const std::string retry_context = std::format(
-      "Name: {} {}\nCity: {}, {}\nPersona: {}", name.first_name, name.last_name,
+       "Name: {} {}\nCity: {}, {}\nPersona: {}", name.first_name,
-      city.location.city, city.location.country, persona.name);
+       name.last_name, city.location.city, city.location.country, persona.name);
 
    constexpr int max_attempts = 3;
    std::string raw;
@@ -51,12 +51,13 @@ UserResult LlamaGenerator::GenerateUser(const EnrichedCity& city,
    int max_tokens = kUserInitialMaxTokens;
 
    for (int attempt = 0; attempt < max_attempts; ++attempt) {
-    raw = this->Infer(system_prompt, user_prompt, max_tokens, kUserJsonGrammar);
+      raw =
+          this->Infer(system_prompt, user_prompt, max_tokens, kUserJsonGrammar);
       if (logger_) {
-      logger_->Log(
+         logger_->Log({.level = LogLevel::Debug,
-          {.level = LogLevel::Debug,
                        .phase = PipelinePhase::UserGeneration,
-           .message = std::format("LlamaGenerator: raw output (attempt {}): {}",
+                       .message = std::format(
+                           "LlamaGenerator: raw output (attempt {}): {}",
                            attempt + 1, raw)});
       }
 
@@ -82,7 +83,8 @@ UserResult LlamaGenerator::GenerateUser(const EnrichedCity& city,
 
       last_error = *validation_error;
       if (logger_) {
-      logger_->Log({.level = LogLevel::Warn,
+         logger_->Log(
+             {.level = LogLevel::Warn,
               .phase = PipelinePhase::UserGeneration,
               .message = std::format(
                   "LlamaGenerator: malformed user JSON (attempt {}): {}",

tooling/pipeline/src/data_generation/llama/helpers.cc

@@ -4,6 +4,8 @@
  * parsing, token decoding, and JSON validation helpers for Llama modules.
  */
 
+#include <llama.h>
+
 #include <algorithm>
 #include <array>
 #include <boost/json.hpp>
@@ -15,8 +17,6 @@
 #include <string_view>
 #include <vector>
 
-#include <llama.h>
-
 #include "data_generation/llama_generator_helpers.h"
 
 namespace {
@@ -82,8 +82,8 @@ bool ReadRequiredTrimmedStringField(const boost::json::object& obj,
 bool HasSchemaPlaceholder(std::span<std::string* const> values) {
    for (const std::string* value : values) {
       std::string lowered = *value;
-    std::ranges::transform(lowered, lowered.begin(),
+      std::ranges::transform(
-                           [](const unsigned char character) {
+          lowered, lowered.begin(), [](const unsigned char character) {
              return static_cast<char>(std::tolower(character));
           });
 
@@ -128,8 +128,8 @@ void AppendTokenPiece(const llama_vocab* vocab, llama_token token,
 
    auto buffer_too_small = [](int32_t result) -> bool { return result < 0; };
 
-  int32_t bytes =
+   int32_t bytes = llama_token_to_piece(vocab, token, buffer.data(),
-      llama_token_to_piece(vocab, token, buffer.data(), buffer.size(), 0, true);
+                                        buffer.size(), 0, true);
 
    if (!buffer_too_small(bytes)) {
       // Append the decoded bytes from the stack buffer.
@@ -190,8 +190,8 @@ std::optional<std::string> ValidateBreweryJson(const std::string& raw,
       return validation_error;
    }
 
-  if (!ReadRequiredTrimmedStringField(obj, "name_local", brewery_out.name_local,
+   if (!ReadRequiredTrimmedStringField(
-                                      &validation_error)) {
+           obj, "name_local", brewery_out.name_local, &validation_error)) {
       return validation_error;
    }
 
@@ -248,7 +248,8 @@ std::optional<std::string> ValidateUserJson(const std::string& raw,
 
    const boost::json::value* activity_weight_field =
        obj.if_contains("activity_weight");
-  if (activity_weight_field == nullptr || !activity_weight_field->is_number()) {
+   if (activity_weight_field == nullptr ||
+       !activity_weight_field->is_number()) {
       return "Missing or invalid numeric field: activity_weight";
    }
 

tooling/pipeline/src/data_generation/llama/infer.cc

@@ -5,6 +5,8 @@
  * output tokens back to text for system+user chat prompts.
  */
 
+#include <llama.h>
+
 #include <algorithm>
 #include <chrono>
 #include <format>
@@ -14,8 +16,6 @@
 #include <string_view>
 #include <vector>
 
-#include <llama.h>
-
 #include "data_generation/llama_generator.h"
 #include "data_generation/llama_generator_helpers.h"
 
@@ -58,7 +58,8 @@ SamplerHandle MakeSamplerChain(const llama_vocab* vocab,
 
    if (!grammar.empty()) {
       const std::string grammar_text(grammar);
-    add_sampler(llama_sampler_init_grammar(vocab, grammar_text.c_str(), "root"),
+      add_sampler(
+          llama_sampler_init_grammar(vocab, grammar_text.c_str(), "root"),
           "LlamaGenerator: failed to initialize grammar sampler");
    }
 
@@ -173,7 +174,8 @@ std::string LlamaGenerator::InferFormatted(const std::string& formatted_prompt,
    prompt_tokens.resize(static_cast<size_t>(token_count));
    if (token_count > prompt_budget) {
       if (logger_) {
-      logger_->Log({.level = LogLevel::Warn,
+         logger_->Log(
+             {.level = LogLevel::Warn,
               .phase = PipelinePhase::BreweryAndBeerGeneration,
               .message = std::format(
                   "LlamaGenerator: prompt too long ({} tokens), "

tooling/pipeline/src/data_generation/llama/llama_generator.cc

@@ -5,14 +5,14 @@
 
 #include "data_generation/llama_generator.h"
 
+#include <llama.h>
+
 #include <filesystem>
 #include <memory>
 #include <random>
 #include <stdexcept>
 #include <string>
 
-#include <llama.h>
-
 #include "data_model/models.h"
 
 static constexpr uint32_t kMaxContextSize = 32768U;

tooling/pipeline/src/data_generation/llama/load.cc

@@ -4,15 +4,15 @@
  * context, and resets prior resources during model initialization.
  */
 
+#include <ggml-backend.h>
+#include <llama.h>
+
 #include <algorithm>
 #include <chrono>
 #include <stdexcept>
 #include <string>
 #include <utility>
 
-#include <ggml-backend.h>
-#include <llama.h>
-
 #include "data_generation/llama_generator.h"
 
 // Maximum batch size for decode operations. Capping the batch prevents