synth-ai 0.2.8.dev4__py3-none-any.whl → 0.2.23.dev3__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (889) hide show
  1. examples/README.md +1 -0
  2. examples/__init__.py +16 -0
  3. examples/analyze_semantic_words.sh +17 -0
  4. examples/baseline/banking77_baseline.py +243 -0
  5. examples/baseline/banking77_pipeline_baseline.py +294 -0
  6. examples/baseline/crafter_baseline.py +407 -0
  7. examples/baseline/pokemon_red_baseline.py +326 -0
  8. examples/baseline/simple_baseline.py +56 -0
  9. examples/baseline/warming_up_to_rl_baseline.py +239 -0
  10. examples/blog_posts/gepa/README.md +355 -0
  11. examples/blog_posts/gepa/configs/banking77_gepa_local.toml +95 -0
  12. examples/blog_posts/gepa/configs/banking77_gepa_test.toml +80 -0
  13. examples/blog_posts/gepa/configs/banking77_mipro_local.toml +50 -0
  14. examples/blog_posts/gepa/configs/banking77_pipeline_gepa_local.toml +101 -0
  15. examples/blog_posts/gepa/configs/banking77_pipeline_gepa_test.toml +96 -0
  16. examples/blog_posts/gepa/configs/hotpotqa_gepa_local.toml +57 -0
  17. examples/blog_posts/gepa/configs/hotpotqa_gepa_qwen.toml +35 -0
  18. examples/blog_posts/gepa/configs/hotpotqa_mipro_local.toml +51 -0
  19. examples/blog_posts/gepa/configs/hover_gepa_local.toml +57 -0
  20. examples/blog_posts/gepa/configs/hover_gepa_qwen.toml +35 -0
  21. examples/blog_posts/gepa/configs/hover_mipro_local.toml +51 -0
  22. examples/blog_posts/gepa/configs/ifbench_gepa_local.toml +57 -0
  23. examples/blog_posts/gepa/configs/ifbench_gepa_qwen.toml +35 -0
  24. examples/blog_posts/gepa/configs/ifbench_mipro_local.toml +51 -0
  25. examples/blog_posts/gepa/configs/pupa_gepa_local.toml +58 -0
  26. examples/blog_posts/gepa/configs/pupa_mipro_local.toml +52 -0
  27. examples/blog_posts/gepa/deploy_banking77_task_app.sh +54 -0
  28. examples/blog_posts/gepa/gepa_baseline.py +204 -0
  29. examples/blog_posts/gepa/query_prompts_example.py +97 -0
  30. examples/blog_posts/gepa/run_gepa_banking77.sh +112 -0
  31. examples/blog_posts/gepa/run_gepa_banking77_pipeline.sh +163 -0
  32. examples/blog_posts/gepa/task_apps.py +105 -0
  33. examples/blog_posts/gepa/test_gepa_local.sh +67 -0
  34. examples/blog_posts/gepa/verify_banking77_setup.sh +123 -0
  35. examples/blog_posts/mipro/README.md +415 -0
  36. examples/blog_posts/mipro/configs/banking77_mipro_local.toml +91 -0
  37. examples/blog_posts/mipro/configs/banking77_mipro_test.toml +87 -0
  38. examples/blog_posts/mipro/configs/banking77_pipeline_mipro_gemini_flash_lite_local.toml +98 -0
  39. examples/blog_posts/mipro/configs/banking77_pipeline_mipro_gpt41mini_local.toml +96 -0
  40. examples/blog_posts/mipro/configs/banking77_pipeline_mipro_local.toml +94 -0
  41. examples/blog_posts/mipro/configs/banking77_pipeline_mipro_test.toml +170 -0
  42. examples/blog_posts/mipro/deploy_banking77_pipeline_task_app.sh +59 -0
  43. examples/blog_posts/mipro/deploy_banking77_task_app.sh +41 -0
  44. examples/blog_posts/mipro/multi_step.md +79 -0
  45. examples/blog_posts/mipro/run_mipro_banking77.sh +191 -0
  46. examples/blog_posts/mipro/run_mipro_banking77_pipeline.sh +171 -0
  47. examples/blog_posts/mipro/run_mipro_banking77_pipeline_gemini_flash_lite.sh +177 -0
  48. examples/blog_posts/mipro/run_mipro_banking77_pipeline_gpt41mini.sh +173 -0
  49. examples/blog_posts/mipro/verify_banking77_setup.sh +117 -0
  50. examples/blog_posts/pokemon_vl/README.md +98 -0
  51. examples/blog_posts/pokemon_vl/configs/eval_gpt5nano.toml +26 -0
  52. examples/blog_posts/pokemon_vl/configs/eval_qwen3_vl.toml +27 -0
  53. examples/blog_posts/pokemon_vl/configs/eval_rl_final.toml +24 -0
  54. examples/blog_posts/pokemon_vl/configs/filter_high_reward.toml +10 -0
  55. examples/blog_posts/pokemon_vl/configs/train_rl_from_sft.toml +43 -0
  56. examples/blog_posts/pokemon_vl/configs/train_sft_qwen4b_vl.toml +40 -0
  57. examples/blog_posts/pokemon_vl/extract_images.py +239 -0
  58. examples/blog_posts/pokemon_vl/pokemon_vl_baseline.py +326 -0
  59. examples/blog_posts/pokemon_vl/run_eval_extract_images.py +209 -0
  60. examples/blog_posts/pokemon_vl/run_qwen_eval_extract_images.py +212 -0
  61. examples/blog_posts/pokemon_vl/text_box_analysis.md +106 -0
  62. examples/blog_posts/warming_up_to_rl/ARCHITECTURE.md +195 -0
  63. examples/blog_posts/warming_up_to_rl/FINAL_TEST_RESULTS.md +127 -0
  64. examples/blog_posts/warming_up_to_rl/INFERENCE_SUCCESS.md +132 -0
  65. examples/blog_posts/warming_up_to_rl/README.md +158 -0
  66. examples/blog_posts/warming_up_to_rl/SMOKE_TESTING.md +164 -0
  67. examples/blog_posts/warming_up_to_rl/SMOKE_TEST_COMPLETE.md +253 -0
  68. examples/blog_posts/warming_up_to_rl/configs/eval_baseline_qwen32b_10x20.toml +25 -0
  69. examples/blog_posts/warming_up_to_rl/configs/eval_ft_qwen4b.toml +25 -0
  70. examples/blog_posts/warming_up_to_rl/configs/eval_ft_qwen4b_10x20.toml +26 -0
  71. examples/blog_posts/warming_up_to_rl/configs/eval_groq_qwen32b.toml +25 -0
  72. examples/blog_posts/warming_up_to_rl/configs/eval_openai_gpt_oss_120b.toml +29 -0
  73. examples/blog_posts/warming_up_to_rl/configs/filter_high_reward_dataset.toml +10 -0
  74. examples/blog_posts/warming_up_to_rl/configs/smoke_test.toml +75 -0
  75. examples/blog_posts/warming_up_to_rl/configs/train_rl_from_sft.toml +91 -0
  76. examples/blog_posts/warming_up_to_rl/configs/train_sft_qwen4b.toml +40 -0
  77. examples/blog_posts/warming_up_to_rl/warming_up_to_rl_baseline.py +187 -0
  78. examples/crafter_debug_render.py +186 -0
  79. examples/dev/qwen3_32b_qlora_4xh100.toml +45 -0
  80. examples/gepa/banking77_pipeline_gepa.toml +96 -0
  81. examples/gepa/multi_stage_gepa_example.toml +84 -0
  82. examples/gepa/run_gepa_banking77_pipeline.sh +157 -0
  83. examples/multi_step/SFT_README.md +147 -0
  84. examples/multi_step/configs/README_verilog_rl.md +77 -0
  85. examples/multi_step/configs/VERILOG_REWARDS.md +103 -0
  86. examples/multi_step/configs/VERILOG_RL_CHECKLIST.md +196 -0
  87. examples/multi_step/configs/crafter_eval_synth_qwen4b.toml +35 -0
  88. examples/multi_step/configs/crafter_eval_text_only_groq_qwen32b.toml +36 -0
  89. examples/multi_step/configs/crafter_rl_outcome.toml +75 -0
  90. examples/multi_step/configs/crafter_rl_stepwise_hosted_judge.toml +145 -0
  91. examples/multi_step/configs/crafter_rl_stepwise_shaped.toml +84 -0
  92. examples/multi_step/configs/crafter_rl_stepwise_simple.toml +79 -0
  93. examples/multi_step/configs/crafter_rl_stepwise_simple_NEW_FORMAT.toml +105 -0
  94. examples/multi_step/configs/crafter_sft_qwen30b_lora.toml +62 -0
  95. examples/multi_step/configs/crafter_synth_backend.md +40 -0
  96. examples/multi_step/configs/verilog_eval_groq_qwen32b.toml +31 -0
  97. examples/multi_step/configs/verilog_eval_synth_qwen8b.toml +33 -0
  98. examples/multi_step/configs/verilog_rl_lora.toml +147 -0
  99. examples/multi_step/convert_traces_to_sft.py +84 -0
  100. examples/multi_step/crafter_rl_lora.md +70 -0
  101. examples/multi_step/judges/crafter_backend_judge.py +220 -0
  102. examples/multi_step/judges/verilog_backend_judge.py +234 -0
  103. examples/multi_step/readme.md +48 -0
  104. examples/multi_step/run_sft_qwen30b.sh +45 -0
  105. examples/multi_step/sse_metrics_streaming_notes.md +357 -0
  106. examples/multi_step/task_app_config_notes.md +494 -0
  107. examples/multi_step/verilog_rl_lora.md +218 -0
  108. examples/qwen_coder/README.md +102 -0
  109. examples/qwen_coder/_shared.py +113 -0
  110. examples/qwen_coder/configs/coder_lora_30b.toml +60 -0
  111. examples/qwen_coder/configs/coder_lora_4b.toml +61 -0
  112. examples/qwen_coder/configs/coder_lora_small.toml +57 -0
  113. examples/qwen_coder/generate_dataset.py +98 -0
  114. examples/qwen_coder/infer_ft_smoke.py +65 -0
  115. examples/qwen_coder/infer_prod_proxy.py +73 -0
  116. examples/qwen_coder/infer_via_synth.py +87 -0
  117. examples/qwen_coder/scripts/infer_coder.sh +19 -0
  118. examples/qwen_coder/scripts/train_coder_30b.sh +22 -0
  119. examples/qwen_coder/sft_full_17b.py +103 -0
  120. examples/qwen_coder/sft_lora_30b.py +110 -0
  121. examples/qwen_coder/subset_jsonl.py +39 -0
  122. examples/qwen_coder/todos.md +38 -0
  123. examples/qwen_coder/validate_jsonl.py +60 -0
  124. examples/qwen_vl/BUGS_AND_FIXES.md +232 -0
  125. examples/qwen_vl/IMAGE_VALIDATION_COMPLETE.md +271 -0
  126. examples/qwen_vl/IMAGE_VALIDATION_SUMMARY.md +260 -0
  127. examples/qwen_vl/INFERENCE_SFT_TESTS.md +412 -0
  128. examples/qwen_vl/NEXT_STEPS_2B.md +325 -0
  129. examples/qwen_vl/QUICKSTART.md +327 -0
  130. examples/qwen_vl/QUICKSTART_RL_VISION.md +110 -0
  131. examples/qwen_vl/README.md +152 -0
  132. examples/qwen_vl/RL_VISION_COMPLETE.md +475 -0
  133. examples/qwen_vl/RL_VISION_TESTING.md +333 -0
  134. examples/qwen_vl/SDK_VISION_INTEGRATION.md +328 -0
  135. examples/qwen_vl/SETUP_COMPLETE.md +274 -0
  136. examples/qwen_vl/VISION_TESTS_COMPLETE.md +489 -0
  137. examples/qwen_vl/VLM_PIPELINE_COMPLETE.md +242 -0
  138. examples/qwen_vl/__init__.py +2 -0
  139. examples/qwen_vl/collect_data_via_cli.md +415 -0
  140. examples/qwen_vl/collect_vision_traces.py +368 -0
  141. examples/qwen_vl/configs/crafter_rl_vision_qwen3vl4b.toml +110 -0
  142. examples/qwen_vl/configs/crafter_vlm_sft_example.toml +59 -0
  143. examples/qwen_vl/configs/eval_gpt4o_mini_vision.toml +26 -0
  144. examples/qwen_vl/configs/eval_gpt4o_vision_proper.toml +29 -0
  145. examples/qwen_vl/configs/eval_gpt5nano_vision.toml +26 -0
  146. examples/qwen_vl/configs/eval_qwen3vl_vision.toml +26 -0
  147. examples/qwen_vl/configs/filter_qwen3vl_sft.toml +49 -0
  148. examples/qwen_vl/configs/filter_vision_sft.toml +52 -0
  149. examples/qwen_vl/configs/filter_vision_test.toml +8 -0
  150. examples/qwen_vl/configs/sft_qwen3_vl_2b_test.toml +54 -0
  151. examples/qwen_vl/crafter_gpt5nano_agent.py +308 -0
  152. examples/qwen_vl/crafter_qwen_vl_agent.py +300 -0
  153. examples/qwen_vl/run_vision_comparison.sh +61 -0
  154. examples/qwen_vl/run_vision_sft_pipeline.sh +175 -0
  155. examples/qwen_vl/test_image_validation.py +201 -0
  156. examples/qwen_vl/test_sft_vision_data.py +110 -0
  157. examples/rl/README.md +169 -0
  158. examples/rl/configs/eval_base_qwen.toml +17 -0
  159. examples/rl/configs/eval_rl_qwen.toml +13 -0
  160. examples/rl/configs/rl_from_base_qwen.toml +62 -0
  161. examples/rl/configs/rl_from_base_qwen17.toml +80 -0
  162. examples/rl/configs/rl_from_ft_qwen.toml +37 -0
  163. examples/rl/download_dataset.py +80 -0
  164. examples/rl/run_eval.py +436 -0
  165. examples/rl/run_rl_and_save.py +111 -0
  166. examples/rl/task_app/README.md +21 -0
  167. examples/rl/task_app/math_single_step.py +990 -0
  168. examples/rl/task_app/math_task_app.py +111 -0
  169. examples/run_crafter_demo.sh +10 -0
  170. examples/sdk_prompt_learning_example.py +55 -0
  171. examples/sft/README.md +139 -0
  172. examples/sft/configs/crafter_fft_qwen0p6b.toml +49 -0
  173. examples/sft/configs/crafter_lora_qwen0p6b.toml +49 -0
  174. examples/sft/evaluate.py +117 -0
  175. examples/sft/export_dataset.py +120 -0
  176. examples/sft/generate_traces.py +164 -0
  177. examples/swe/__init__.py +12 -0
  178. examples/swe/task_app/README.md +135 -0
  179. examples/swe/task_app/__init__.py +2 -0
  180. examples/swe/task_app/grpo_swe_mini.py +604 -0
  181. examples/swe/task_app/grpo_swe_mini_task_app.py +124 -0
  182. examples/swe/task_app/hosted/README.md +173 -0
  183. examples/swe/task_app/hosted/__init__.py +5 -0
  184. examples/swe/task_app/hosted/branching.py +143 -0
  185. examples/swe/task_app/hosted/environment_routes.py +1289 -0
  186. examples/swe/task_app/hosted/envs/__init__.py +1 -0
  187. examples/swe/task_app/hosted/envs/crafter/__init__.py +6 -0
  188. examples/swe/task_app/hosted/envs/crafter/app.py +1 -0
  189. examples/swe/task_app/hosted/envs/crafter/environment.py +522 -0
  190. examples/swe/task_app/hosted/envs/crafter/policy.py +478 -0
  191. examples/swe/task_app/hosted/envs/crafter/react_agent.py +108 -0
  192. examples/swe/task_app/hosted/envs/crafter/shared.py +305 -0
  193. examples/swe/task_app/hosted/envs/crafter/tools.py +47 -0
  194. examples/swe/task_app/hosted/envs/mini_swe/__init__.py +8 -0
  195. examples/swe/task_app/hosted/envs/mini_swe/environment.py +1191 -0
  196. examples/swe/task_app/hosted/envs/mini_swe/policy.py +355 -0
  197. examples/swe/task_app/hosted/envs/mini_swe/shared.py +83 -0
  198. examples/swe/task_app/hosted/envs/mini_swe/tools.py +96 -0
  199. examples/swe/task_app/hosted/hosted_app.py +204 -0
  200. examples/swe/task_app/hosted/inference/__init__.py +5 -0
  201. examples/swe/task_app/hosted/inference/openai_client.py +584 -0
  202. examples/swe/task_app/hosted/main.py +100 -0
  203. examples/swe/task_app/hosted/policy_routes.py +1094 -0
  204. examples/swe/task_app/hosted/registry.py +195 -0
  205. examples/swe/task_app/hosted/rollout.py +1905 -0
  206. examples/swe/task_app/hosted/storage/__init__.py +5 -0
  207. examples/swe/task_app/hosted/storage/volume.py +211 -0
  208. examples/swe/task_app/hosted/test_agents.py +161 -0
  209. examples/swe/task_app/hosted/test_service.py +136 -0
  210. examples/swe/task_app/hosted/utils.py +62 -0
  211. examples/swe/task_app/morph_backend.py +178 -0
  212. examples/task_apps/IMAGE_ONLY_EVAL_QUICKSTART.md +258 -0
  213. examples/task_apps/TESTING.md +275 -0
  214. examples/task_apps/banking77/__init__.py +6 -0
  215. examples/task_apps/banking77/banking77_task_app.py +912 -0
  216. examples/task_apps/banking77/deploy_wrapper.py +46 -0
  217. examples/task_apps/banking77_pipeline/__init__.py +6 -0
  218. examples/task_apps/banking77_pipeline/banking77_pipeline_task_app.py +489 -0
  219. examples/task_apps/banking77_pipeline/deploy_wrapper.py +50 -0
  220. examples/task_apps/crafter/CREATE_SFT_DATASET.md +286 -0
  221. examples/task_apps/crafter/EVAL_IMAGE_ONLY_RESULTS.md +152 -0
  222. examples/task_apps/crafter/FILTER_COMMAND_STATUS.md +187 -0
  223. examples/task_apps/crafter/FILTER_COMMAND_SUCCESS.md +281 -0
  224. examples/task_apps/crafter/QUERY_EXAMPLES.md +203 -0
  225. examples/task_apps/crafter/README_IMAGE_ONLY_EVAL.md +316 -0
  226. examples/task_apps/crafter/eval_image_only_gpt4o.toml +28 -0
  227. examples/task_apps/crafter/eval_text_only_groq_llama.toml +36 -0
  228. examples/task_apps/crafter/filter_sft_dataset.toml +16 -0
  229. examples/task_apps/crafter/task_app/README.md +42 -0
  230. examples/task_apps/crafter/task_app/__init__.py +5 -0
  231. examples/task_apps/crafter/task_app/grpo_crafter.py +1055 -0
  232. examples/task_apps/crafter/task_app/grpo_crafter_task_app.py +146 -0
  233. examples/task_apps/crafter/task_app/synth_envs_hosted/README.md +173 -0
  234. examples/task_apps/crafter/task_app/synth_envs_hosted/__init__.py +5 -0
  235. examples/task_apps/crafter/task_app/synth_envs_hosted/branching.py +143 -0
  236. examples/task_apps/crafter/task_app/synth_envs_hosted/environment_routes.py +1226 -0
  237. examples/task_apps/crafter/task_app/synth_envs_hosted/envs/__init__.py +1 -0
  238. examples/task_apps/crafter/task_app/synth_envs_hosted/envs/crafter/__init__.py +6 -0
  239. examples/task_apps/crafter/task_app/synth_envs_hosted/envs/crafter/app.py +1 -0
  240. examples/task_apps/crafter/task_app/synth_envs_hosted/envs/crafter/environment.py +532 -0
  241. examples/task_apps/crafter/task_app/synth_envs_hosted/envs/crafter/policy.py +583 -0
  242. examples/task_apps/crafter/task_app/synth_envs_hosted/envs/crafter/react_agent.py +122 -0
  243. examples/task_apps/crafter/task_app/synth_envs_hosted/envs/crafter/shared.py +305 -0
  244. examples/task_apps/crafter/task_app/synth_envs_hosted/envs/crafter/tools.py +47 -0
  245. examples/task_apps/crafter/task_app/synth_envs_hosted/hosted_app.py +253 -0
  246. examples/task_apps/crafter/task_app/synth_envs_hosted/inference/__init__.py +5 -0
  247. examples/task_apps/crafter/task_app/synth_envs_hosted/inference/openai_client.py +999 -0
  248. examples/task_apps/crafter/task_app/synth_envs_hosted/main.py +100 -0
  249. examples/task_apps/crafter/task_app/synth_envs_hosted/policy_routes.py +1252 -0
  250. examples/task_apps/crafter/task_app/synth_envs_hosted/registry.py +195 -0
  251. examples/task_apps/crafter/task_app/synth_envs_hosted/rollout.py +2233 -0
  252. examples/task_apps/crafter/task_app/synth_envs_hosted/storage/__init__.py +5 -0
  253. examples/task_apps/crafter/task_app/synth_envs_hosted/storage/volume.py +211 -0
  254. examples/task_apps/crafter/task_app/synth_envs_hosted/test_agents.py +161 -0
  255. examples/task_apps/crafter/task_app/synth_envs_hosted/test_service.py +136 -0
  256. examples/task_apps/crafter/task_app/synth_envs_hosted/utils.py +411 -0
  257. examples/task_apps/dev/pokemon_emerald/__init__.py +2 -0
  258. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/README.md +811 -0
  259. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/agent/__init__.py +120 -0
  260. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/agent/action.py +160 -0
  261. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/agent/memory.py +155 -0
  262. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/agent/perception.py +69 -0
  263. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/agent/planning.py +96 -0
  264. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/agent/simple.py +1502 -0
  265. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/agent/system_prompt.py +4 -0
  266. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/grab_map.py +68 -0
  267. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/manual.py +216 -0
  268. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/pokemon_env/__init__.py +35 -0
  269. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/pokemon_env/emerald_utils.py +631 -0
  270. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/pokemon_env/emulator.py +1544 -0
  271. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/pokemon_env/enums.py +1428 -0
  272. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/pokemon_env/memory_reader.py +4848 -0
  273. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/pokemon_env/types.py +41 -0
  274. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/pokemon_env/utils.py +298 -0
  275. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/pyproject.toml +95 -0
  276. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/run.py +204 -0
  277. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/server/app.py +2152 -0
  278. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/server/client.py +429 -0
  279. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/server/frame_server.py +155 -0
  280. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/README.md +78 -0
  281. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/run_tests.py +122 -0
  282. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/test_agent_direct.py +76 -0
  283. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/test_agent_prompts.py +413 -0
  284. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/test_battle_state_formatting.py +204 -0
  285. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/test_dialogue_detection.py +133 -0
  286. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/test_dialogue_detection_comprehensive.py +229 -0
  287. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/test_direct_agent_emulator.py +300 -0
  288. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/test_fps_adjustment_pytest.py +205 -0
  289. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/test_house_to_outside_direct.py +200 -0
  290. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/test_house_to_outside_transition.py +284 -0
  291. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/test_map_ground_truth_comparison.py +468 -0
  292. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/test_memory_map.py +575 -0
  293. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/test_server_map_validation.py +311 -0
  294. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests/test_torchic_state.py +259 -0
  295. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/utils/anticheat.py +372 -0
  296. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/utils/checkpoint.py +296 -0
  297. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/utils/error_handler.py +275 -0
  298. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/utils/get_local_ip.py +22 -0
  299. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/utils/helpers.py +44 -0
  300. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/utils/llm_logger.py +514 -0
  301. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/utils/map_formatter.py +415 -0
  302. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/utils/map_stitcher.py +1763 -0
  303. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/utils/map_stitcher_singleton.py +33 -0
  304. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/utils/map_trimmer.py +106 -0
  305. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/utils/map_visualizer.py +334 -0
  306. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/utils/ocr_dialogue.py +1020 -0
  307. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/utils/recording.py +188 -0
  308. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/utils/state_formatter.py +1481 -0
  309. examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/utils/vlm.py +862 -0
  310. examples/task_apps/dev/pokemon_emerald/modal_app.py +114 -0
  311. examples/task_apps/dev/pokemon_emerald/task_app/README.md +81 -0
  312. examples/task_apps/dev/pokemon_emerald/task_app/__init__.py +6 -0
  313. examples/task_apps/dev/pokemon_emerald/task_app/pokemon_emerald.py +685 -0
  314. examples/task_apps/enron/__init__.py +2 -0
  315. examples/task_apps/enron/eval_groq_qwen32.toml +16 -0
  316. examples/task_apps/enron/filter_sft.toml +5 -0
  317. examples/task_apps/enron/task_app/README.md +14 -0
  318. examples/task_apps/enron/task_app/__init__.py +1 -0
  319. examples/task_apps/enron/task_app/grpo_enron.py +906 -0
  320. examples/task_apps/enron/task_app/grpo_enron_task_app.py +146 -0
  321. examples/task_apps/enron/tests/__init__.py +4 -0
  322. examples/task_apps/enron/tests/conftest.py +115 -0
  323. examples/task_apps/enron/tests/integration/__init__.py +4 -0
  324. examples/task_apps/enron/tests/integration/test_enron_eval.py +179 -0
  325. examples/task_apps/enron/tests/integration/test_enron_rollout.py +135 -0
  326. examples/task_apps/enron/tests/unit/__init__.py +4 -0
  327. examples/task_apps/enron/tests/unit/test_enron_environment.py +126 -0
  328. examples/task_apps/gepa_benchmarks/__init__.py +7 -0
  329. examples/task_apps/gepa_benchmarks/common.py +260 -0
  330. examples/task_apps/gepa_benchmarks/hotpotqa_task_app.py +507 -0
  331. examples/task_apps/gepa_benchmarks/hover_task_app.py +436 -0
  332. examples/task_apps/gepa_benchmarks/ifbench_task_app.py +563 -0
  333. examples/task_apps/gepa_benchmarks/pupa_task_app.py +460 -0
  334. examples/task_apps/math/README.md +21 -0
  335. examples/task_apps/math/math_single_step.py +1000 -0
  336. examples/task_apps/math/math_task_app.py +115 -0
  337. examples/task_apps/pokemon_battle/__init__.py +2 -0
  338. examples/task_apps/pokemon_battle/modal_app.py +104 -0
  339. examples/task_apps/pokemon_battle/task_app/README.md +68 -0
  340. examples/task_apps/pokemon_battle/task_app/__init__.py +6 -0
  341. examples/task_apps/pokemon_battle/task_app/pokemon_showdown.py +932 -0
  342. examples/task_apps/pokemon_red/EVAL_IMAGE_ONLY_COMPLETE.md +283 -0
  343. examples/task_apps/pokemon_red/EVAL_IMAGE_ONLY_STATUS.md +155 -0
  344. examples/task_apps/pokemon_red/README.md +356 -0
  345. examples/task_apps/pokemon_red/README_IMAGE_ONLY_EVAL.md +428 -0
  346. examples/task_apps/pokemon_red/__init__.py +3 -0
  347. examples/task_apps/pokemon_red/eval_image_only_gpt4o.toml +30 -0
  348. examples/task_apps/pokemon_red/eval_pokemon_red_policy.py +224 -0
  349. examples/task_apps/pokemon_red/pallet_town_rl_config.toml +75 -0
  350. examples/task_apps/pokemon_red/task_app.py +1048 -0
  351. examples/task_apps/pokemon_red/test_pallet_town_rewards.py +193 -0
  352. examples/task_apps/sokoban/README.md +306 -0
  353. examples/task_apps/sokoban/__init__.py +3 -0
  354. examples/task_apps/sokoban/eval_groq_qwen32.toml +16 -0
  355. examples/task_apps/sokoban/eval_openai_gpt5.toml +16 -0
  356. examples/task_apps/sokoban/filter_sft.toml +5 -0
  357. examples/task_apps/sokoban/task_app.py +1058 -0
  358. examples/task_apps/sokoban/tests/__init__.py +4 -0
  359. examples/task_apps/sokoban/tests/conftest.py +113 -0
  360. examples/task_apps/sokoban/tests/integration/__init__.py +4 -0
  361. examples/task_apps/sokoban/tests/integration/test_sokoban_eval.py +57 -0
  362. examples/task_apps/sokoban/tests/integration/test_sokoban_rollout.py +198 -0
  363. examples/task_apps/sokoban/tests/unit/__init__.py +4 -0
  364. examples/task_apps/sokoban/tests/unit/test_sokoban_environment.py +114 -0
  365. examples/task_apps/verilog/__init__.py +1 -0
  366. examples/task_apps/verilog/eval_groq_qwen32b.toml +22 -0
  367. examples/task_apps/verilog/filter_sft.toml +5 -0
  368. examples/task_apps/verilog/task_app/README.md +12 -0
  369. examples/task_apps/verilog/task_app/__init__.py +1 -0
  370. examples/task_apps/verilog/task_app/grpo_verilog.py +1166 -0
  371. examples/task_apps/verilog/task_app/grpo_verilog_task_app.py +145 -0
  372. examples/task_apps/verilog/tests/__init__.py +4 -0
  373. examples/task_apps/verilog/tests/conftest.py +115 -0
  374. examples/task_apps/verilog/tests/integration/__init__.py +4 -0
  375. examples/task_apps/verilog/tests/integration/test_verilog_eval.py +181 -0
  376. examples/task_apps/verilog/tests/integration/test_verilog_rollout.py +55 -0
  377. examples/task_apps/verilog/tests/unit/__init__.py +4 -0
  378. examples/task_apps/verilog/tests/unit/test_verilog_scoring.py +118 -0
  379. examples/tunnel_gepa_banking77/README.md +106 -0
  380. examples/tunnel_gepa_banking77/banking77_gepa_tunnel.toml +95 -0
  381. examples/tunnel_gepa_banking77/keep_tunnel_running.py +60 -0
  382. examples/tunnel_gepa_banking77/run_gepa_with_tunnel.sh +226 -0
  383. examples/vlm/PROPOSAL.md +53 -0
  384. examples/vlm/README.md +68 -0
  385. examples/vlm/configs/crafter_vlm_gpt4o.toml +49 -0
  386. examples/vlm/crafter_image_only_agent.py +207 -0
  387. examples/vlm/crafter_openai_vlm_agent.py +275 -0
  388. examples/vlm/filter_image_rows.py +63 -0
  389. examples/vlm/run_crafter_vlm_benchmark.py +316 -0
  390. examples/warming_up_to_rl/_utils.py +92 -0
  391. examples/warming_up_to_rl/analyze_trace_db.py +422 -0
  392. examples/warming_up_to_rl/configs/crafter_fft.toml +53 -0
  393. examples/warming_up_to_rl/configs/crafter_fft_4b.toml +54 -0
  394. examples/warming_up_to_rl/configs/eval_fft_qwen4b.toml +22 -0
  395. examples/warming_up_to_rl/configs/eval_groq_qwen32b.toml +15 -0
  396. examples/warming_up_to_rl/configs/eval_modal_qwen4b.toml +24 -0
  397. examples/warming_up_to_rl/configs/eval_stepwise_complex.toml +35 -0
  398. examples/warming_up_to_rl/configs/eval_stepwise_consistent.toml +26 -0
  399. examples/warming_up_to_rl/configs/eval_stepwise_per_achievement.toml +36 -0
  400. examples/warming_up_to_rl/configs/eval_stepwise_simple.toml +32 -0
  401. examples/warming_up_to_rl/configs/rl_from_base_qwen4b.toml +85 -0
  402. examples/warming_up_to_rl/configs/rl_from_ft.toml +58 -0
  403. examples/warming_up_to_rl/export_trace_sft.py +837 -0
  404. examples/warming_up_to_rl/groq_test.py +97 -0
  405. examples/warming_up_to_rl/manage_secrets.py +131 -0
  406. examples/warming_up_to_rl/old/event_rewards.md +234 -0
  407. examples/warming_up_to_rl/old/notes.md +73 -0
  408. examples/warming_up_to_rl/readme.md +110 -0
  409. examples/warming_up_to_rl/run_eval.py +736 -0
  410. examples/warming_up_to_rl/run_fft_and_save.py +380 -0
  411. examples/warming_up_to_rl/run_local_rollout.py +239 -0
  412. examples/warming_up_to_rl/run_local_rollout_modal.py +248 -0
  413. examples/warming_up_to_rl/run_local_rollout_parallel.py +405 -0
  414. examples/warming_up_to_rl/run_local_rollout_traced.py +477 -0
  415. examples/warming_up_to_rl/run_rl_and_save.py +124 -0
  416. examples/warming_up_to_rl/run_rollout_remote.py +156 -0
  417. examples/warming_up_to_rl/task_app/README.md +42 -0
  418. examples/warming_up_to_rl/task_app/grpo_crafter.py +876 -0
  419. examples/warming_up_to_rl/task_app/grpo_crafter_task_app.py +135 -0
  420. examples/warming_up_to_rl/task_app/synth_envs_hosted/README.md +173 -0
  421. examples/warming_up_to_rl/task_app/synth_envs_hosted/__init__.py +5 -0
  422. examples/warming_up_to_rl/task_app/synth_envs_hosted/branching.py +143 -0
  423. examples/warming_up_to_rl/task_app/synth_envs_hosted/environment_routes.py +1226 -0
  424. examples/warming_up_to_rl/task_app/synth_envs_hosted/envs/__init__.py +1 -0
  425. examples/warming_up_to_rl/task_app/synth_envs_hosted/envs/crafter/__init__.py +6 -0
  426. examples/warming_up_to_rl/task_app/synth_envs_hosted/envs/crafter/app.py +1 -0
  427. examples/warming_up_to_rl/task_app/synth_envs_hosted/envs/crafter/environment.py +522 -0
  428. examples/warming_up_to_rl/task_app/synth_envs_hosted/envs/crafter/policy.py +454 -0
  429. examples/warming_up_to_rl/task_app/synth_envs_hosted/envs/crafter/react_agent.py +108 -0
  430. examples/warming_up_to_rl/task_app/synth_envs_hosted/envs/crafter/shared.py +305 -0
  431. examples/warming_up_to_rl/task_app/synth_envs_hosted/envs/crafter/tools.py +47 -0
  432. examples/warming_up_to_rl/task_app/synth_envs_hosted/hosted_app.py +253 -0
  433. examples/warming_up_to_rl/task_app/synth_envs_hosted/inference/__init__.py +5 -0
  434. examples/warming_up_to_rl/task_app/synth_envs_hosted/inference/openai_client.py +729 -0
  435. examples/warming_up_to_rl/task_app/synth_envs_hosted/main.py +100 -0
  436. examples/warming_up_to_rl/task_app/synth_envs_hosted/policy_routes.py +1114 -0
  437. examples/warming_up_to_rl/task_app/synth_envs_hosted/registry.py +195 -0
  438. examples/warming_up_to_rl/task_app/synth_envs_hosted/rollout.py +1891 -0
  439. examples/warming_up_to_rl/task_app/synth_envs_hosted/storage/__init__.py +5 -0
  440. examples/warming_up_to_rl/task_app/synth_envs_hosted/storage/volume.py +211 -0
  441. examples/warming_up_to_rl/task_app/synth_envs_hosted/test_agents.py +161 -0
  442. examples/warming_up_to_rl/task_app/synth_envs_hosted/test_service.py +137 -0
  443. examples/warming_up_to_rl/task_app/synth_envs_hosted/utils.py +129 -0
  444. examples/workflows/math_rl/configs/eval_base_qwen.toml +15 -0
  445. examples/workflows/math_rl/configs/eval_rl_qwen.toml +11 -0
  446. examples/workflows/math_rl/configs/rl_from_base_qwen.toml +62 -0
  447. examples/workflows/math_rl/configs/rl_from_base_qwen17.toml +80 -0
  448. examples/workflows/math_rl/configs/rl_from_ft_qwen.toml +35 -0
  449. examples/workflows/math_rl/download_dataset.py +80 -0
  450. examples/workflows/math_rl/run_eval.py +436 -0
  451. examples/workflows/math_rl/run_rl_and_save.py +111 -0
  452. synth_ai/__init__.py +47 -23
  453. synth_ai/_utils/__init__.py +47 -0
  454. synth_ai/_utils/base_url.py +10 -0
  455. synth_ai/_utils/http.py +10 -0
  456. synth_ai/_utils/prompts.py +10 -0
  457. synth_ai/_utils/task_app_state.py +12 -0
  458. synth_ai/_utils/user_config.py +10 -0
  459. synth_ai/api/models/supported.py +514 -0
  460. synth_ai/api/train/__init__.py +63 -0
  461. synth_ai/api/train/builders.py +473 -0
  462. synth_ai/api/train/cli.py +1185 -0
  463. synth_ai/api/train/config_finder.py +246 -0
  464. synth_ai/api/train/configs/__init__.py +65 -0
  465. synth_ai/api/train/configs/prompt_learning.py +496 -0
  466. synth_ai/api/train/configs/rl.py +188 -0
  467. synth_ai/api/train/configs/sft.py +99 -0
  468. synth_ai/api/train/configs/shared.py +81 -0
  469. synth_ai/api/train/env_resolver.py +352 -0
  470. synth_ai/api/train/pollers.py +91 -0
  471. synth_ai/api/train/prompt_learning.py +425 -0
  472. synth_ai/api/train/sft.py +390 -0
  473. synth_ai/api/train/supported_algos.py +147 -0
  474. synth_ai/api/train/task_app.py +195 -0
  475. synth_ai/api/train/utils.py +244 -0
  476. synth_ai/api/train/validators.py +1117 -0
  477. synth_ai/api/tunnel.py +49 -0
  478. synth_ai/auth/credentials.py +94 -0
  479. synth_ai/baseline/__init__.py +25 -0
  480. synth_ai/baseline/config.py +209 -0
  481. synth_ai/baseline/discovery.py +214 -0
  482. synth_ai/baseline/execution.py +146 -0
  483. synth_ai/cfgs.py +227 -0
  484. synth_ai/cli/__init__.py +90 -45
  485. synth_ai/cli/_modal_wrapper.py +31 -0
  486. synth_ai/cli/_storage.py +20 -0
  487. synth_ai/cli/_typer_patch.py +47 -0
  488. synth_ai/cli/_validate_task_app.py +29 -0
  489. synth_ai/cli/balance.py +16 -4
  490. synth_ai/cli/calc.py +36 -21
  491. synth_ai/cli/claude.py +70 -0
  492. synth_ai/cli/codex.py +267 -0
  493. synth_ai/cli/commands/__init__.py +18 -0
  494. synth_ai/cli/commands/baseline/__init__.py +12 -0
  495. synth_ai/cli/commands/baseline/core.py +637 -0
  496. synth_ai/cli/commands/baseline/list.py +93 -0
  497. synth_ai/cli/commands/demo/__init__.py +6 -0
  498. synth_ai/cli/commands/demo/core.py +163 -0
  499. synth_ai/cli/commands/eval/__init__.py +19 -0
  500. synth_ai/cli/commands/eval/core.py +1112 -0
  501. synth_ai/cli/commands/eval/errors.py +81 -0
  502. synth_ai/cli/commands/eval/validation.py +133 -0
  503. synth_ai/cli/commands/filter/__init__.py +12 -0
  504. synth_ai/cli/commands/filter/core.py +424 -0
  505. synth_ai/cli/commands/filter/errors.py +55 -0
  506. synth_ai/cli/commands/filter/validation.py +77 -0
  507. synth_ai/cli/commands/help/__init__.py +185 -0
  508. synth_ai/cli/commands/help/core.py +72 -0
  509. synth_ai/cli/commands/smoke/__init__.py +7 -0
  510. synth_ai/cli/commands/smoke/core.py +1437 -0
  511. synth_ai/cli/commands/status/__init__.py +66 -0
  512. synth_ai/cli/commands/status/client.py +192 -0
  513. synth_ai/cli/commands/status/config.py +92 -0
  514. synth_ai/cli/commands/status/errors.py +20 -0
  515. synth_ai/cli/commands/status/formatters.py +164 -0
  516. synth_ai/cli/commands/status/subcommands/__init__.py +9 -0
  517. synth_ai/cli/commands/status/subcommands/files.py +79 -0
  518. synth_ai/cli/commands/status/subcommands/jobs.py +334 -0
  519. synth_ai/cli/commands/status/subcommands/models.py +79 -0
  520. synth_ai/cli/commands/status/subcommands/pricing.py +22 -0
  521. synth_ai/cli/commands/status/subcommands/runs.py +81 -0
  522. synth_ai/cli/commands/status/subcommands/session.py +183 -0
  523. synth_ai/cli/commands/status/subcommands/summary.py +47 -0
  524. synth_ai/cli/commands/status/subcommands/usage.py +203 -0
  525. synth_ai/cli/commands/status/utils.py +114 -0
  526. synth_ai/cli/commands/train/__init__.py +53 -0
  527. synth_ai/cli/commands/train/core.py +21 -0
  528. synth_ai/cli/commands/train/errors.py +117 -0
  529. synth_ai/cli/commands/train/judge_schemas.py +200 -0
  530. synth_ai/cli/commands/train/judge_validation.py +305 -0
  531. synth_ai/cli/commands/train/validation.py +386 -0
  532. synth_ai/cli/demo.py +32 -140
  533. synth_ai/cli/deploy.py +233 -0
  534. synth_ai/cli/eval/__init__.py +36 -0
  535. synth_ai/cli/eval/core.py +5 -0
  536. synth_ai/cli/eval/errors.py +31 -0
  537. synth_ai/cli/eval/validation.py +5 -0
  538. synth_ai/cli/filter/__init__.py +28 -0
  539. synth_ai/cli/filter/core.py +5 -0
  540. synth_ai/cli/filter/errors.py +23 -0
  541. synth_ai/cli/filter/validation.py +5 -0
  542. synth_ai/cli/legacy_root_backup.py +28 -22
  543. synth_ai/cli/lib/__init__.py +10 -0
  544. synth_ai/cli/lib/task_app_discovery.py +7 -0
  545. synth_ai/cli/lib/task_app_env.py +518 -0
  546. synth_ai/cli/mcp.py +34 -0
  547. synth_ai/cli/modal_serve/__init__.py +12 -0
  548. synth_ai/cli/modal_serve/core.py +14 -0
  549. synth_ai/cli/modal_serve/errors.py +8 -0
  550. synth_ai/cli/modal_serve/validation.py +11 -0
  551. synth_ai/cli/opencode.py +256 -0
  552. synth_ai/cli/recent.py +13 -7
  553. synth_ai/cli/rl_demo.py +166 -114
  554. synth_ai/cli/root.py +143 -112
  555. synth_ai/cli/serve/__init__.py +12 -0
  556. synth_ai/cli/serve/core.py +14 -0
  557. synth_ai/cli/serve/errors.py +8 -0
  558. synth_ai/cli/serve/validation.py +11 -0
  559. synth_ai/cli/setup.py +49 -0
  560. synth_ai/cli/status.py +7 -125
  561. synth_ai/cli/task_app_deploy.py +7 -0
  562. synth_ai/cli/task_app_list.py +25 -0
  563. synth_ai/cli/task_app_modal_serve.py +11 -0
  564. synth_ai/cli/task_app_serve.py +11 -0
  565. synth_ai/cli/task_apps.py +3134 -0
  566. synth_ai/cli/traces.py +9 -5
  567. synth_ai/cli/train/__init__.py +12 -0
  568. synth_ai/cli/train/core.py +21 -0
  569. synth_ai/cli/train/errors.py +8 -0
  570. synth_ai/cli/train/validation.py +24 -0
  571. synth_ai/cli/train.py +5 -0
  572. synth_ai/cli/turso.py +73 -0
  573. synth_ai/cli/watch.py +13 -18
  574. synth_ai/demos/__init__.py +10 -0
  575. synth_ai/demos/core/__init__.py +28 -1
  576. synth_ai/demos/core/cli.py +745 -416
  577. synth_ai/demos/crafter/__init__.py +1 -0
  578. synth_ai/demos/crafter/crafter_fft_4b.toml +55 -0
  579. synth_ai/demos/crafter/grpo_crafter_task_app.py +185 -0
  580. synth_ai/demos/crafter/rl_from_base_qwen4b.toml +74 -0
  581. synth_ai/demos/demo_registry.py +176 -0
  582. synth_ai/demos/demo_task_apps/__init__.py +7 -1
  583. synth_ai/demos/demo_task_apps/core.py +75 -37
  584. synth_ai/demos/demo_task_apps/crafter/__init__.py +1 -0
  585. synth_ai/demos/demo_task_apps/crafter/configs/crafter_fft_4b.toml +53 -0
  586. synth_ai/demos/demo_task_apps/crafter/configs/rl_from_base_qwen4b.toml +73 -0
  587. synth_ai/demos/demo_task_apps/crafter/grpo_crafter_task_app.py +184 -0
  588. synth_ai/demos/demo_task_apps/math/_common.py +1 -2
  589. synth_ai/demos/demo_task_apps/math/app.py +2 -1
  590. synth_ai/demos/demo_task_apps/math/config.toml +55 -110
  591. synth_ai/demos/demo_task_apps/math/deploy_modal.py +3 -6
  592. synth_ai/demos/demo_task_apps/math/modal_task_app.py +491 -166
  593. synth_ai/demos/demo_task_apps/math/task_app_entry.py +37 -0
  594. synth_ai/demos/math/__init__.py +1 -0
  595. synth_ai/demos/math/_common.py +16 -0
  596. synth_ai/demos/math/app.py +38 -0
  597. synth_ai/demos/math/config.toml +76 -0
  598. synth_ai/demos/math/deploy_modal.py +54 -0
  599. synth_ai/demos/math/modal_task_app.py +703 -0
  600. synth_ai/demos/math/task_app_entry.py +51 -0
  601. synth_ai/environments/environment/core.py +7 -1
  602. synth_ai/environments/examples/bandit/engine.py +12 -5
  603. synth_ai/environments/examples/bandit/environment.py +0 -1
  604. synth_ai/environments/examples/bandit/taskset.py +4 -4
  605. synth_ai/environments/examples/crafter_classic/engine_deterministic_patch.py +7 -4
  606. synth_ai/environments/examples/crafter_classic/engine_serialization_patch_v3.py +9 -5
  607. synth_ai/environments/examples/crafter_classic/environment.py +93 -2
  608. synth_ai/environments/examples/crafter_classic/world_config_patch_simple.py +4 -3
  609. synth_ai/environments/examples/enron/engine.py +7 -2
  610. synth_ai/environments/examples/enron/environment.py +68 -0
  611. synth_ai/environments/examples/red/engine.py +60 -12
  612. synth_ai/environments/examples/red/engine_helpers/memory_map.py +7 -0
  613. synth_ai/environments/examples/red/engine_helpers/reward_components.py +151 -179
  614. synth_ai/environments/examples/red/engine_helpers/reward_library/pallet_town_progression.py +477 -0
  615. synth_ai/environments/examples/red/engine_helpers/state_extraction.py +32 -0
  616. synth_ai/environments/examples/red/environment.py +86 -0
  617. synth_ai/environments/examples/red/trace_hooks_v3.py +168 -0
  618. synth_ai/environments/examples/sokoban/taskset.py +116 -0
  619. synth_ai/environments/examples/verilog/engine.py +104 -12
  620. synth_ai/environments/examples/wordle/environment.py +0 -1
  621. synth_ai/environments/reproducibility/tree.py +5 -6
  622. synth_ai/environments/service/app.py +11 -12
  623. synth_ai/environments/service/core_routes.py +10 -9
  624. synth_ai/environments/stateful/engine.py +1 -1
  625. synth_ai/environments/tasks/core.py +1 -0
  626. synth_ai/environments/tasks/filters.py +5 -6
  627. synth_ai/environments/tasks/utils.py +4 -5
  628. synth_ai/evals/__init__.py +15 -0
  629. synth_ai/evals/base.py +14 -5
  630. synth_ai/evals/client.py +82 -0
  631. synth_ai/evals/types.py +42 -0
  632. synth_ai/http.py +8 -22
  633. synth_ai/http_client.py +45 -12
  634. synth_ai/inference/__init__.py +0 -2
  635. synth_ai/inference/client.py +21 -7
  636. synth_ai/jobs/client.py +129 -80
  637. synth_ai/judge_schemas.py +127 -0
  638. synth_ai/learning/__init__.py +51 -6
  639. synth_ai/learning/algorithms.py +14 -0
  640. synth_ai/learning/client.py +122 -30
  641. synth_ai/learning/config.py +2 -40
  642. synth_ai/learning/constants.py +0 -2
  643. synth_ai/learning/ft_client.py +4 -56
  644. synth_ai/learning/health.py +14 -8
  645. synth_ai/learning/jobs.py +43 -47
  646. synth_ai/learning/prompt_learning_client.py +276 -0
  647. synth_ai/learning/prompt_learning_types.py +185 -0
  648. synth_ai/{rl → learning/rl}/__init__.py +14 -5
  649. synth_ai/learning/rl/client.py +269 -0
  650. synth_ai/learning/rl/config.py +31 -0
  651. synth_ai/{rl → learning/rl}/contracts.py +5 -10
  652. synth_ai/{rl → learning/rl}/env_keys.py +45 -16
  653. synth_ai/learning/rl/secrets.py +13 -0
  654. synth_ai/learning/rl_client.py +2 -253
  655. synth_ai/learning/sft/__init__.py +29 -0
  656. synth_ai/learning/sft/client.py +68 -0
  657. synth_ai/learning/sft/config.py +270 -0
  658. synth_ai/learning/sft/data.py +698 -0
  659. synth_ai/learning/sse.py +25 -26
  660. synth_ai/learning/validators.py +29 -25
  661. synth_ai/mcp/__init__.py +5 -0
  662. synth_ai/mcp/__main__.py +8 -0
  663. synth_ai/mcp/main.py +254 -0
  664. synth_ai/mcp/setup.py +100 -0
  665. synth_ai/modal.py +257 -0
  666. synth_ai/pricing/__init__.py +3 -0
  667. synth_ai/pricing/model_pricing.py +64 -0
  668. synth_ai/session/__init__.py +75 -0
  669. synth_ai/session/client.py +383 -0
  670. synth_ai/session/constants.py +63 -0
  671. synth_ai/session/exceptions.py +105 -0
  672. synth_ai/session/manager.py +139 -0
  673. synth_ai/session/models.py +89 -0
  674. synth_ai/session/query.py +110 -0
  675. synth_ai/spec/__init__.py +46 -0
  676. synth_ai/spec/dataclasses.py +149 -0
  677. synth_ai/spec/loader.py +144 -0
  678. synth_ai/spec/serializer.py +199 -0
  679. synth_ai/spec/validation.py +250 -0
  680. synth_ai/streaming/__init__.py +29 -0
  681. synth_ai/streaming/config.py +94 -0
  682. synth_ai/streaming/handlers.py +589 -0
  683. synth_ai/streaming/streamer.py +320 -0
  684. synth_ai/streaming/types.py +95 -0
  685. synth_ai/task/__init__.py +116 -3
  686. synth_ai/task/apps/__init__.py +132 -0
  687. synth_ai/task/auth.py +165 -0
  688. synth_ai/task/client.py +167 -0
  689. synth_ai/task/config.py +261 -0
  690. synth_ai/task/contracts.py +173 -57
  691. synth_ai/task/datasets.py +108 -0
  692. synth_ai/task/errors.py +50 -0
  693. synth_ai/task/health.py +17 -11
  694. synth_ai/task/inference_api.py +101 -0
  695. synth_ai/task/json.py +111 -0
  696. synth_ai/task/proxy.py +251 -0
  697. synth_ai/task/rubrics/__init__.py +55 -0
  698. synth_ai/task/rubrics/loaders.py +156 -0
  699. synth_ai/task/rubrics/models.py +57 -0
  700. synth_ai/task/rubrics/scoring.py +116 -0
  701. synth_ai/task/rubrics/strict.py +149 -0
  702. synth_ai/task/rubrics.py +219 -0
  703. synth_ai/task/server.py +432 -0
  704. synth_ai/task/trace_correlation_helpers.py +328 -0
  705. synth_ai/task/tracing_utils.py +95 -0
  706. synth_ai/task/validators.py +449 -6
  707. synth_ai/task/vendors.py +59 -0
  708. synth_ai/tracing_v3/__init__.py +4 -0
  709. synth_ai/tracing_v3/abstractions.py +21 -4
  710. synth_ai/tracing_v3/config.py +167 -22
  711. synth_ai/tracing_v3/constants.py +21 -0
  712. synth_ai/tracing_v3/db_config.py +42 -29
  713. synth_ai/tracing_v3/decorators.py +80 -45
  714. synth_ai/tracing_v3/examples/basic_usage.py +15 -9
  715. synth_ai/tracing_v3/hooks.py +6 -4
  716. synth_ai/tracing_v3/llm_call_record_helpers.py +161 -61
  717. synth_ai/tracing_v3/migration_helper.py +1 -2
  718. synth_ai/tracing_v3/replica_sync.py +12 -7
  719. synth_ai/tracing_v3/serialization.py +130 -0
  720. synth_ai/tracing_v3/session_tracer.py +86 -21
  721. synth_ai/tracing_v3/storage/base.py +98 -12
  722. synth_ai/tracing_v3/storage/config.py +63 -16
  723. synth_ai/tracing_v3/storage/factory.py +11 -9
  724. synth_ai/tracing_v3/storage/utils.py +15 -11
  725. synth_ai/tracing_v3/trace_utils.py +317 -0
  726. synth_ai/tracing_v3/turso/__init__.py +8 -21
  727. synth_ai/tracing_v3/turso/daemon.py +123 -15
  728. synth_ai/tracing_v3/turso/models.py +5 -2
  729. synth_ai/tracing_v3/turso/native_manager.py +1293 -0
  730. synth_ai/tracing_v3/utils.py +5 -4
  731. synth_ai/tunnel.py +143 -0
  732. synth_ai/tunnel_deploy.py +278 -0
  733. synth_ai/types.py +8 -0
  734. synth_ai/urls.py +11 -0
  735. synth_ai/utils/__init__.py +166 -0
  736. synth_ai/utils/agents.py +74 -0
  737. synth_ai/utils/apps.py +152 -0
  738. synth_ai/utils/base_url.py +94 -0
  739. synth_ai/utils/bin.py +39 -0
  740. synth_ai/utils/claude.py +36 -0
  741. synth_ai/utils/cli.py +284 -0
  742. synth_ai/utils/config.py +81 -0
  743. synth_ai/utils/env.py +346 -0
  744. synth_ai/utils/errors.py +85 -0
  745. synth_ai/utils/http.py +172 -0
  746. synth_ai/utils/json.py +72 -0
  747. synth_ai/utils/log_filter.py +99 -0
  748. synth_ai/utils/logging.py +198 -0
  749. synth_ai/utils/modal.py +299 -0
  750. synth_ai/utils/paths.py +95 -0
  751. synth_ai/utils/process.py +233 -0
  752. synth_ai/utils/prompts.py +39 -0
  753. synth_ai/utils/sqld.py +122 -0
  754. synth_ai/utils/ssl.py +25 -0
  755. synth_ai/utils/task_app_discovery.py +882 -0
  756. synth_ai/utils/task_app_env.py +186 -0
  757. synth_ai/utils/task_app_state.py +318 -0
  758. synth_ai/utils/tunnel/__init__.py +12 -0
  759. synth_ai/utils/tunnel/config.py +55 -0
  760. synth_ai/utils/user_config.py +137 -0
  761. synth_ai/uvicorn.py +77 -0
  762. synth_ai-0.2.23.dev3.dist-info/METADATA +357 -0
  763. synth_ai-0.2.23.dev3.dist-info/RECORD +983 -0
  764. {synth_ai-0.2.8.dev4.dist-info → synth_ai-0.2.23.dev3.dist-info}/entry_points.txt +0 -1
  765. {synth_ai-0.2.8.dev4.dist-info → synth_ai-0.2.23.dev3.dist-info}/top_level.txt +1 -0
  766. synth_ai/cli/man.py +0 -106
  767. synth_ai/core/experiment.py +0 -15
  768. synth_ai/core/system.py +0 -15
  769. synth_ai/environments/examples/sokoban/units/astar_common.py +0 -95
  770. synth_ai/experimental/synth_oss.py +0 -446
  771. synth_ai/handshake.py +0 -63
  772. synth_ai/install_sqld.sh +0 -40
  773. synth_ai/learning/offline/dpo.py +0 -0
  774. synth_ai/learning/offline/providers.py +0 -7
  775. synth_ai/learning/offline/sft.py +0 -0
  776. synth_ai/learning/offline/shared.py +0 -0
  777. synth_ai/learning/online/grpo.py +0 -0
  778. synth_ai/learning/online/irft.py +0 -0
  779. synth_ai/learning/prompts/banking77_injection_eval.py +0 -168
  780. synth_ai/learning/prompts/gepa.py +0 -0
  781. synth_ai/learning/prompts/hello_world_in_context_injection_ex.py +0 -213
  782. synth_ai/learning/prompts/mipro.py +0 -289
  783. synth_ai/learning/prompts/random_search.py +0 -246
  784. synth_ai/learning/prompts/run_mipro_banking77.py +0 -172
  785. synth_ai/learning/prompts/run_random_search_banking77.py +0 -324
  786. synth_ai/lm/__init__.py +0 -51
  787. synth_ai/lm/caching/constants.py +0 -6
  788. synth_ai/lm/caching/dbs.py +0 -0
  789. synth_ai/lm/caching/ephemeral.py +0 -102
  790. synth_ai/lm/caching/handler.py +0 -137
  791. synth_ai/lm/caching/initialize.py +0 -11
  792. synth_ai/lm/caching/persistent.py +0 -114
  793. synth_ai/lm/config.py +0 -110
  794. synth_ai/lm/constants.py +0 -32
  795. synth_ai/lm/core/__init__.py +0 -8
  796. synth_ai/lm/core/all.py +0 -73
  797. synth_ai/lm/core/exceptions.py +0 -7
  798. synth_ai/lm/core/main.py +0 -319
  799. synth_ai/lm/core/main_v3.py +0 -594
  800. synth_ai/lm/core/synth_models.py +0 -48
  801. synth_ai/lm/core/vendor_clients.py +0 -188
  802. synth_ai/lm/cost/monitor.py +0 -1
  803. synth_ai/lm/cost/statefulness.py +0 -1
  804. synth_ai/lm/injection.py +0 -80
  805. synth_ai/lm/overrides.py +0 -206
  806. synth_ai/lm/provider_support/__init__.py +0 -8
  807. synth_ai/lm/provider_support/anthropic.py +0 -972
  808. synth_ai/lm/provider_support/openai.py +0 -1139
  809. synth_ai/lm/provider_support/suppress_logging.py +0 -31
  810. synth_ai/lm/structured_outputs/handler.py +0 -440
  811. synth_ai/lm/structured_outputs/inject.py +0 -297
  812. synth_ai/lm/structured_outputs/rehabilitate.py +0 -185
  813. synth_ai/lm/tools/__init__.py +0 -3
  814. synth_ai/lm/tools/base.py +0 -172
  815. synth_ai/lm/unified_interface.py +0 -202
  816. synth_ai/lm/vendors/base.py +0 -81
  817. synth_ai/lm/vendors/core/anthropic_api.py +0 -387
  818. synth_ai/lm/vendors/core/gemini_api.py +0 -292
  819. synth_ai/lm/vendors/core/mistral_api.py +0 -322
  820. synth_ai/lm/vendors/core/openai_api.py +0 -225
  821. synth_ai/lm/vendors/core/synth_dev_api.py +0 -0
  822. synth_ai/lm/vendors/local/ollama.py +0 -0
  823. synth_ai/lm/vendors/openai_standard.py +0 -780
  824. synth_ai/lm/vendors/openai_standard_responses.py +0 -256
  825. synth_ai/lm/vendors/retries.py +0 -22
  826. synth_ai/lm/vendors/supported/custom_endpoint.py +0 -417
  827. synth_ai/lm/vendors/supported/deepseek.py +0 -69
  828. synth_ai/lm/vendors/supported/grok.py +0 -75
  829. synth_ai/lm/vendors/supported/groq.py +0 -16
  830. synth_ai/lm/vendors/supported/ollama.py +0 -15
  831. synth_ai/lm/vendors/supported/openrouter.py +0 -74
  832. synth_ai/lm/vendors/supported/together.py +0 -11
  833. synth_ai/lm/vendors/synth_client.py +0 -808
  834. synth_ai/lm/warmup.py +0 -186
  835. synth_ai/rl/secrets.py +0 -19
  836. synth_ai/scripts/verify_rewards.py +0 -100
  837. synth_ai/tracing/__init__.py +0 -30
  838. synth_ai/tracing_v1/__init__.py +0 -33
  839. synth_ai/tracing_v3/turso/manager.py +0 -760
  840. synth_ai/v0/tracing/abstractions.py +0 -224
  841. synth_ai/v0/tracing/base_client.py +0 -91
  842. synth_ai/v0/tracing/client_manager.py +0 -131
  843. synth_ai/v0/tracing/config.py +0 -142
  844. synth_ai/v0/tracing/context.py +0 -146
  845. synth_ai/v0/tracing/decorators.py +0 -682
  846. synth_ai/v0/tracing/events/__init__.py +0 -0
  847. synth_ai/v0/tracing/events/manage.py +0 -147
  848. synth_ai/v0/tracing/events/scope.py +0 -86
  849. synth_ai/v0/tracing/events/store.py +0 -228
  850. synth_ai/v0/tracing/immediate_client.py +0 -151
  851. synth_ai/v0/tracing/local.py +0 -18
  852. synth_ai/v0/tracing/log_client_base.py +0 -73
  853. synth_ai/v0/tracing/retry_queue.py +0 -186
  854. synth_ai/v0/tracing/trackers.py +0 -515
  855. synth_ai/v0/tracing/upload.py +0 -512
  856. synth_ai/v0/tracing/utils.py +0 -9
  857. synth_ai/v0/tracing_v1/__init__.py +0 -16
  858. synth_ai/v0/tracing_v1/abstractions.py +0 -224
  859. synth_ai/v0/tracing_v1/base_client.py +0 -91
  860. synth_ai/v0/tracing_v1/client_manager.py +0 -131
  861. synth_ai/v0/tracing_v1/config.py +0 -142
  862. synth_ai/v0/tracing_v1/context.py +0 -146
  863. synth_ai/v0/tracing_v1/decorators.py +0 -703
  864. synth_ai/v0/tracing_v1/events/__init__.py +0 -0
  865. synth_ai/v0/tracing_v1/events/manage.py +0 -147
  866. synth_ai/v0/tracing_v1/events/scope.py +0 -86
  867. synth_ai/v0/tracing_v1/events/store.py +0 -228
  868. synth_ai/v0/tracing_v1/immediate_client.py +0 -151
  869. synth_ai/v0/tracing_v1/local.py +0 -18
  870. synth_ai/v0/tracing_v1/log_client_base.py +0 -73
  871. synth_ai/v0/tracing_v1/retry_queue.py +0 -186
  872. synth_ai/v0/tracing_v1/trackers.py +0 -515
  873. synth_ai/v0/tracing_v1/upload.py +0 -527
  874. synth_ai/v0/tracing_v1/utils.py +0 -9
  875. synth_ai/zyk/__init__.py +0 -30
  876. synth_ai-0.2.8.dev4.dist-info/METADATA +0 -129
  877. synth_ai-0.2.8.dev4.dist-info/RECORD +0 -420
  878. {synth_ai/lm/caching → examples/task_apps}/__init__.py +0 -0
  879. {synth_ai/lm/cost → examples/task_apps/crafter}/__init__.py +0 -0
  880. {synth_ai/lm/structured_outputs → examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/server}/__init__.py +0 -0
  881. {synth_ai/lm/vendors → examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/tests}/__init__.py +0 -0
  882. {synth_ai/lm/vendors/core → examples/task_apps/dev/pokemon_emerald/external/pokeagent-speedrun/utils}/__init__.py +0 -0
  883. {synth_ai/lm/vendors/local → examples/task_apps/math}/__init__.py +0 -0
  884. {synth_ai/lm/vendors/supported → examples/workflows}/__init__.py +0 -0
  885. {synth_ai/v0/tracing → examples/workflows/math_rl}/__init__.py +0 -0
  886. /synth_ai/{compound/cais.py → cli/__main__.py} +0 -0
  887. /synth_ai/{learning/filtering.py → py.typed} +0 -0
  888. {synth_ai-0.2.8.dev4.dist-info → synth_ai-0.2.23.dev3.dist-info}/WHEEL +0 -0
  889. {synth_ai-0.2.8.dev4.dist-info → synth_ai-0.2.23.dev3.dist-info}/licenses/LICENSE +0 -0
@@ -1,527 +0,0 @@
1
- import asyncio
2
- import json
3
- import logging
4
- import os
5
- import ssl
6
- import time
7
- import warnings
8
- from typing import Any, TypedDict
9
-
10
- import requests
11
- from dotenv import load_dotenv
12
- from pydantic import BaseModel, ConfigDict, field_validator
13
- from requests.adapters import HTTPAdapter
14
- from urllib3.poolmanager import PoolManager
15
-
16
- from synth_ai.config.base_url import PROD_BASE_URL_DEFAULT
17
-
18
- from .abstractions import Dataset, SystemTrace
19
- from .events.store import event_store
20
-
21
- load_dotenv()
22
-
23
-
24
- # Issue deprecation warning and raise not supported error for backend upload functionality
25
- def _raise_not_supported_error():
26
- """Raise an error for backend upload functionality that is no longer supported."""
27
- warnings.warn(
28
- "Backend upload functionality in synth_ai.tracing_v1 is no longer supported. "
29
- "Please use synth_ai.tracing_v2 for new tracing functionality.",
30
- DeprecationWarning,
31
- stacklevel=3,
32
- )
33
- raise NotImplementedError(
34
- "Backend upload functionality is no longer supported in synth_ai.tracing_v1. "
35
- "Please use synth_ai.tracing_v2 instead."
36
- )
37
-
38
-
39
- # NOTE: This may cause memory issues in the future
40
- def validate_json(data: dict[str, Any]) -> None:
41
- """Validate that a dictionary contains only JSON-serializable values.
42
-
43
- Args:
44
- data: Dictionary to validate for JSON serialization
45
-
46
- Raises:
47
- ValueError: If the dictionary contains non-serializable values
48
- """
49
-
50
- try:
51
- json.dumps(data)
52
- except (TypeError, OverflowError) as e:
53
- raise ValueError(f"Contains non-JSON-serializable values: {e}. {data}")
54
-
55
-
56
- def createPayload(dataset: Dataset, traces: list[SystemTrace]) -> dict[str, Any]:
57
- payload = {
58
- "traces": [trace.to_dict() for trace in traces], # Convert SystemTrace objects to dicts
59
- "dataset": dataset.to_dict(),
60
- }
61
- return payload
62
-
63
-
64
- class TLSAdapter(HTTPAdapter):
65
- def init_poolmanager(self, connections, maxsize, block=False):
66
- """Create and initialize the urllib3 PoolManager."""
67
- ctx = ssl.create_default_context()
68
- ctx.set_ciphers("DEFAULT@SECLEVEL=1")
69
- self.poolmanager = PoolManager(
70
- num_pools=connections,
71
- maxsize=maxsize,
72
- block=block,
73
- ssl_version=ssl.PROTOCOL_TLSv1_2,
74
- ssl_context=ctx,
75
- )
76
-
77
-
78
- def load_signed_url(signed_url: str, dataset: Dataset, traces: list[SystemTrace]) -> None:
79
- payload = createPayload(dataset, traces)
80
- validate_json(payload)
81
-
82
- session = requests.Session()
83
- adapter = TLSAdapter()
84
- session.mount("https://", adapter)
85
-
86
- try:
87
- response = session.put(
88
- signed_url, json=payload, headers={"Content-Type": "application/json"}
89
- )
90
- response.raise_for_status()
91
- except requests.exceptions.RequestException as e:
92
- print(f"Error making request: {str(e)}")
93
- print(f"Request payload: {payload}") # Debugging info
94
- raise
95
-
96
- if response.status_code != 200:
97
- raise ValueError(
98
- f"Failed to load signed URL Status Code: {response.status_code} "
99
- f"Response: {response.text}, Signed URL: {signed_url}"
100
- )
101
- else:
102
- pass
103
-
104
-
105
- def send_system_traces_s3(
106
- dataset: Dataset,
107
- traces: list[SystemTrace],
108
- base_url: str,
109
- api_key: str,
110
- system_id: str,
111
- system_name: str,
112
- verbose: bool = False,
113
- ):
114
- upload_id, signed_url = get_upload_id(base_url, api_key, system_id, system_name, verbose)
115
- load_signed_url(signed_url, dataset, traces)
116
-
117
- token_url = f"{base_url}/v1/auth/token"
118
- try:
119
- token_response = requests.get(token_url, headers={"customer_specific_api_key": api_key})
120
- token_response.raise_for_status()
121
- access_token = token_response.json()["access_token"]
122
- except requests.exceptions.RequestException as e:
123
- logging.error(f"Error obtaining access token: {e}")
124
- raise
125
-
126
- api_url = f"{base_url}/v1/uploads/process-upload/{upload_id}"
127
- data = {"signed_url": signed_url}
128
- headers = {
129
- "Content-Type": "application/json",
130
- "Authorization": f"Bearer {access_token}",
131
- }
132
-
133
- try:
134
- response = requests.post(api_url, headers=headers, json=data)
135
- response.raise_for_status()
136
-
137
- response_data = response.json()
138
- upload_id = response_data.get("upload_id")
139
- signed_url = response_data.get("signed_url")
140
- return upload_id, signed_url
141
- except requests.exceptions.HTTPError as e:
142
- logging.error(f"HTTP error occurred: {e}")
143
- raise
144
- except Exception as e:
145
- logging.error(f"An error occurred: {e}")
146
- raise
147
-
148
-
149
- def get_upload_id(
150
- base_url: str, api_key: str, system_id: str, system_name: str, verbose: bool = False
151
- ) -> tuple[str, str]:
152
- """
153
- Modified client-side function to send both system_id and system_name.
154
- """
155
- token_url = f"{base_url}/v1/auth/token"
156
- token_response = requests.get(token_url, headers={"customer_specific_api_key": api_key})
157
- token_response.raise_for_status()
158
- access_token = token_response.json()["access_token"]
159
-
160
- # Include system_name in the query parameters
161
- api_url = (
162
- f"{base_url}/v1/uploads/get-upload-id-signed-url?"
163
- f"system_id={system_id}&system_name={system_name}"
164
- )
165
- headers = {
166
- "Content-Type": "application/json",
167
- "Authorization": f"Bearer {access_token}",
168
- }
169
-
170
- try:
171
- response = requests.get(api_url, headers=headers)
172
- response.raise_for_status()
173
- upload_id = response.json()["upload_id"]
174
- signed_url = response.json()["signed_url"]
175
- return upload_id, signed_url
176
- except requests.exceptions.HTTPError as e:
177
- logging.error(f"HTTP error occurred: {e}")
178
- raise
179
- except Exception as e:
180
- logging.error(f"An error occurred: {e}")
181
- raise
182
-
183
-
184
- class UploadValidator(BaseModel):
185
- """Validator for upload data format."""
186
-
187
- model_config = ConfigDict(
188
- from_attributes=True, # Replaces the deprecated orm_mode
189
- validate_assignment=True,
190
- extra="forbid", # Prevent additional fields
191
- )
192
-
193
- traces: list[dict[str, Any]]
194
- dataset: dict[str, Any]
195
-
196
- @field_validator("traces")
197
- @classmethod
198
- def validate_traces(cls, traces: list[dict[str, Any]]) -> list[dict[str, Any]]:
199
- if not traces:
200
- raise ValueError("Traces list cannot be empty")
201
-
202
- for trace in traces:
203
- # Validate required fields in each trace
204
- if "system_instance_id" not in trace:
205
- raise ValueError("Each trace must have a system_instance_id")
206
- if "partition" not in trace:
207
- raise ValueError("Each trace must have a partition")
208
-
209
- # Validate metadata if present
210
- if "metadata" in trace and trace["metadata"] is not None:
211
- if not isinstance(trace["metadata"], dict):
212
- raise ValueError("Metadata must be a dictionary")
213
-
214
- # Validate partition structure
215
- partition = trace["partition"]
216
- if not isinstance(partition, list):
217
- raise ValueError("Partition must be a list")
218
-
219
- for part in partition:
220
- if "partition_index" not in part:
221
- raise ValueError("Each partition element must have a partition_index")
222
- if "events" not in part:
223
- raise ValueError("Each partition element must have an events list")
224
-
225
- # Validate events
226
- events = part["events"]
227
- if not isinstance(events, list):
228
- raise ValueError("Events must be a list")
229
-
230
- for event in events:
231
- required_fields = [
232
- "event_type",
233
- "opened",
234
- "closed",
235
- "partition_index",
236
- ]
237
- missing_fields = [f for f in required_fields if f not in event]
238
- if missing_fields:
239
- raise ValueError(f"Event missing required fields: {missing_fields}")
240
-
241
- return traces
242
-
243
- @field_validator("dataset")
244
- @classmethod
245
- def validate_dataset(cls, dataset: dict[str, Any]) -> dict[str, Any]:
246
- required_fields = ["questions", "reward_signals"]
247
- missing_fields = [f for f in required_fields if f not in dataset]
248
- if missing_fields:
249
- raise ValueError(f"Dataset missing required fields: {missing_fields}")
250
-
251
- # Validate questions
252
- questions = dataset["questions"]
253
- if not isinstance(questions, list):
254
- raise ValueError("Questions must be a list")
255
-
256
- for question in questions:
257
- if "intent" not in question or "criteria" not in question:
258
- raise ValueError("Each question must have intent and criteria")
259
-
260
- # Validate reward signals
261
- reward_signals = dataset["reward_signals"]
262
- if not isinstance(reward_signals, list):
263
- raise ValueError("Reward signals must be a list")
264
-
265
- return dataset
266
-
267
-
268
- def validate_upload(traces: list[dict[str, Any]], dataset: dict[str, Any]):
269
- # Validate the upload format before sending to server.
270
- # Raises ValueError if validation fails.
271
- try:
272
- UploadValidator(traces=traces, dataset=dataset)
273
- return True
274
- except ValueError as e:
275
- raise ValueError(f"Upload validation failed: {str(e)}")
276
-
277
-
278
- def is_event_loop_running():
279
- try:
280
- asyncio.get_running_loop() # Check if there's a running event loop
281
- return True
282
- except RuntimeError:
283
- # This exception is raised if no event loop is running
284
- return False
285
-
286
-
287
- def format_upload_output(
288
- dataset: Dataset, traces: list[SystemTrace]
289
- ) -> tuple[list[dict[str, Any]], list[dict[str, Any]], list[dict[str, Any]]]:
290
- # Format questions array
291
- questions_data = [
292
- {"intent": q.intent, "criteria": q.criteria, "id": q.id} for q in dataset.questions
293
- ]
294
-
295
- # Format reward signals array with error handling
296
- reward_signals_data = [
297
- {
298
- "system_instance_id": rs.system_instance_id,
299
- "reward": rs.reward,
300
- "question_id": rs.question_id,
301
- "annotation": rs.annotation if hasattr(rs, "annotation") else None,
302
- }
303
- for rs in dataset.reward_signals
304
- ]
305
-
306
- # Format traces array
307
- traces_data = [
308
- {
309
- "system_instance_id": t.system_instance_id,
310
- "metadata": t.metadata if t.metadata else None,
311
- "partition": [
312
- {
313
- "partition_index": p.partition_index,
314
- "events": [e.to_dict() for e in p.events],
315
- }
316
- for p in t.partition
317
- ],
318
- }
319
- for t in traces
320
- ]
321
-
322
- return questions_data, reward_signals_data, traces_data
323
-
324
-
325
- class UploadIdResponse(TypedDict):
326
- message: str
327
- upload_id: str
328
- signed_url: str
329
-
330
-
331
- class ProcessUploadResponse(TypedDict):
332
- status: str
333
- upload_id: str
334
- signed_url: str
335
-
336
-
337
- def upload(
338
- dataset: Dataset,
339
- traces: list[SystemTrace] = [],
340
- verbose: bool = False,
341
- show_payload: bool = False,
342
- ) -> tuple[
343
- ProcessUploadResponse,
344
- list[dict[str, Any]],
345
- list[dict[str, Any]],
346
- list[dict[str, Any]],
347
- ]:
348
- """Upload all system traces and dataset to the server.
349
-
350
- DEPRECATED: Backend upload functionality is no longer supported in v1.
351
-
352
- Args:
353
- dataset: Dataset containing questions and reward signals
354
- traces: List of system traces to upload
355
- verbose: Whether to print verbose output
356
- show_payload: Whether to show the payload being sent
357
-
358
- Returns:
359
- Tuple containing:
360
- - response: Server response with status, upload_id, and signed_url
361
- - questions_json: List of formatted questions
362
- - reward_signals_json: List of formatted reward signals
363
- - traces_json: List of formatted traces
364
-
365
- Raises:
366
- ValueError: If no system traces found or validation fails
367
- requests.exceptions.HTTPError: If server request fails
368
- RuntimeError: If SYNTH_API_KEY environment variable not set
369
- """
370
- _raise_not_supported_error()
371
- return upload_helper(dataset, traces, verbose, show_payload)
372
-
373
-
374
- def upload_helper(
375
- dataset: Dataset,
376
- traces: list[SystemTrace] = [],
377
- verbose: bool = False,
378
- show_payload: bool = False,
379
- ) -> tuple[
380
- ProcessUploadResponse,
381
- list[dict[str, Any]],
382
- list[dict[str, Any]],
383
- list[dict[str, Any]],
384
- ]:
385
- """Helper function to handle the upload process.
386
-
387
- Returns same type as upload() function.
388
- """
389
- _raise_not_supported_error()
390
- api_key = os.getenv("SYNTH_API_KEY")
391
- if not api_key:
392
- raise ValueError("SYNTH_API_KEY environment variable not set")
393
- base_url = os.getenv("SYNTH_ENDPOINT_OVERRIDE", PROD_BASE_URL_DEFAULT)
394
-
395
- from .decorators import _local, active_events_var
396
- from .trackers import synth_tracker_async
397
-
398
- # First close any tracker events
399
- if hasattr(synth_tracker_async, "active_events"):
400
- for event_type, event in list(synth_tracker_async.active_events.items()):
401
- if event and event.closed is None:
402
- event.closed = time.time()
403
- try:
404
- event_store.add_event(
405
- event.system_name,
406
- event.system_id,
407
- event.system_instance_id,
408
- event,
409
- )
410
- if verbose:
411
- print(f"Closed and stored tracker async event: {event_type}")
412
- except Exception as e:
413
- logging.error(f"Failed to store tracker event {event_type}: {str(e)}")
414
- synth_tracker_async.active_events.clear()
415
-
416
- # End all active events before uploading
417
- if hasattr(_local, "active_events"):
418
- for event_type, event in _local.active_events.items():
419
- if event and event.closed is None:
420
- event.closed = time.time()
421
- if hasattr(_local, "system_instance_id"):
422
- try:
423
- event_store.add_event(
424
- _local.system_name,
425
- _local.system_id,
426
- _local.system_instance_id,
427
- event,
428
- )
429
- if verbose:
430
- print(f"Closed and stored active event: {event_type}")
431
- except Exception as e:
432
- logging.error(f"Failed to store event {event_type}: {str(e)}")
433
- _local.active_events.clear()
434
-
435
- # NEW: Close all open asynchronous events
436
- active_events_async = active_events_var.get()
437
- if active_events_async:
438
- current_time = time.time()
439
- for event_type, event in list(active_events_async.items()):
440
- if event and event.closed is None:
441
- event.closed = current_time
442
- try:
443
- event_store.add_event(
444
- event.system_name,
445
- event.system_id,
446
- event.system_instance_id,
447
- event,
448
- )
449
- if verbose:
450
- print(f"Closed and stored async event: {event_type}")
451
- except Exception as e:
452
- logging.error(f"Failed to store async event {event_type}: {str(e)}")
453
- active_events_var.set({})
454
-
455
- # Also close any unclosed events in existing traces
456
- logged_traces = event_store.get_system_traces()
457
- traces = logged_traces + traces
458
- # traces = event_store.get_system_traces() if len(traces) == 0 else traces
459
- current_time = time.time()
460
- for trace in traces:
461
- for partition in trace.partition:
462
- for event in partition.events:
463
- if event.closed is None:
464
- event.closed = current_time
465
- event_store.add_event(
466
- trace.system_name,
467
- trace.system_id,
468
- trace.system_instance_id,
469
- event,
470
- )
471
- if verbose:
472
- print(f"Closed existing unclosed event: {event.event_type}")
473
-
474
- try:
475
- # Get traces and convert to dict format
476
- if len(traces) == 0:
477
- raise ValueError("No system traces found")
478
- traces_dict = [trace.to_dict() for trace in traces]
479
- dataset_dict = dataset.to_dict()
480
-
481
- # Validate upload format
482
- if verbose:
483
- print("Validating upload format...")
484
- validate_upload(traces_dict, dataset_dict)
485
- if verbose:
486
- print("Upload format validation successful")
487
-
488
- # Send to server
489
- upload_id, signed_url = send_system_traces_s3(
490
- dataset=dataset,
491
- traces=traces,
492
- base_url=base_url,
493
- api_key=api_key,
494
- system_id=traces[0].system_id,
495
- system_name=traces[0].system_name,
496
- verbose=verbose,
497
- )
498
-
499
- questions_json, reward_signals_json, traces_json = format_upload_output(dataset, traces)
500
- return (
501
- {
502
- "status": "success",
503
- "upload_id": upload_id,
504
- "signed_url": signed_url,
505
- },
506
- questions_json,
507
- reward_signals_json,
508
- traces_json,
509
- )
510
-
511
- except ValueError as e:
512
- if verbose:
513
- print("Validation error:", str(e))
514
- print("\nTraces:")
515
- print(json.dumps(traces_dict, indent=2))
516
- print("\nDataset:")
517
- print(json.dumps(dataset_dict, indent=2))
518
- raise
519
- except requests.exceptions.HTTPError as e:
520
- if verbose:
521
- print("HTTP error occurred:", e)
522
- print("\nTraces:")
523
- print(json.dumps(traces_dict, indent=2))
524
- print("\nDataset:")
525
- print(json.dumps(dataset_dict, indent=2))
526
- raise
527
- # ruff: noqa
@@ -1,9 +0,0 @@
1
- import uuid
2
-
3
-
4
- def get_system_id(system_name: str) -> str:
5
- """Create a deterministic system_instance_id from system_name using UUID5."""
6
- if not system_name:
7
- raise ValueError("system_name cannot be empty")
8
- system_id = uuid.uuid5(uuid.NAMESPACE_DNS, system_name)
9
- return str(system_id)
synth_ai/zyk/__init__.py DELETED
@@ -1,30 +0,0 @@
1
- """
2
- DEPRECATED: This module has been moved to synth_ai.lm
3
-
4
- The synth_ai.zyk module is deprecated and will be removed in a future version.
5
- Please update your imports:
6
-
7
- OLD: from synth_ai.zyk import LM
8
- NEW: from synth_ai.lm.core.main import LM
9
-
10
- or
11
-
12
- NEW: from synth_ai import LM # (recommended)
13
- """
14
-
15
- # ruff: noqa: E402
16
- import warnings
17
-
18
- # Issue deprecation warning
19
- warnings.warn(
20
- "synth_ai.zyk is deprecated and will be removed in a future version. "
21
- "Please use 'from synth_ai import LM' or 'from synth_ai.lm.core.main import LM' instead.",
22
- DeprecationWarning,
23
- stacklevel=2,
24
- )
25
-
26
- # Import from new location for backward compatibility
27
- from synth_ai.lm.core.main import LM
28
- from synth_ai.lm.vendors.base import BaseLMResponse
29
-
30
- __all__ = ["LM", "BaseLMResponse"]
@@ -1,129 +0,0 @@
1
- Metadata-Version: 2.4
2
- Name: synth-ai
3
- Version: 0.2.8.dev4
4
- Summary: RL as a service SDK - Core AI functionality and tracing
5
- Author-email: Synth AI <josh@usesynth.ai>
6
- License-Expression: MIT
7
- Project-URL: Homepage, https://github.com/synth-laboratories/synth-ai
8
- Project-URL: Repository, https://github.com/synth-laboratories/synth-ai
9
- Project-URL: Issues, https://github.com/synth-laboratories/synth-ai/issues
10
- Requires-Python: >=3.11
11
- Description-Content-Type: text/markdown
12
- License-File: LICENSE
13
- Requires-Dist: pydantic>=2.0.0
14
- Requires-Dist: python-dotenv>=1.0.1
15
- Requires-Dist: requests>=2.32.3
16
- Requires-Dist: urllib3>=2.3.0
17
- Requires-Dist: certifi>=2024.8.30
18
- Requires-Dist: tqdm>=4.66.4
19
- Requires-Dist: jsonschema>=4.23.0
20
- Requires-Dist: backoff>=2.0.0
21
- Requires-Dist: typing_extensions>=4.0.0
22
- Requires-Dist: openai>=1.99.0
23
- Requires-Dist: anthropic>=0.42.0
24
- Requires-Dist: langfuse<3.0.0,>=2.53.9
25
- Requires-Dist: opentelemetry-api<1.27.0,>=1.26.0
26
- Requires-Dist: opentelemetry-sdk<1.27.0,>=1.26.0
27
- Requires-Dist: diskcache>=5.6.3
28
- Requires-Dist: groq>=0.30.0
29
- Requires-Dist: google-genai>=1.26.0
30
- Requires-Dist: together>=1.5.21
31
- Requires-Dist: mistralai>=1.9.2
32
- Requires-Dist: fastapi>=0.115.12
33
- Requires-Dist: uvicorn>=0.34.2
34
- Requires-Dist: numpy>=2.2.3
35
- Requires-Dist: networkx>=3.4.2
36
- Requires-Dist: redis>=6.2.0
37
- Requires-Dist: duckdb>=1.0.0
38
- Requires-Dist: pandas>=2.2.3
39
- Requires-Dist: ty>=0.0.1a5
40
- Requires-Dist: toml>=0.10.2
41
- Requires-Dist: sqlalchemy>=2.0.42
42
- Requires-Dist: aiosqlite>=0.21.0
43
- Requires-Dist: greenlet>=3.2.3
44
- Requires-Dist: libsql>=0.1.8
45
- Requires-Dist: pynacl>=1.5.0
46
- Requires-Dist: google-api-core>=2.25.1
47
- Requires-Dist: google-generativeai>=0.8.5
48
- Requires-Dist: crafter>=1.8.3
49
- Requires-Dist: click>=8.1.0
50
- Requires-Dist: textual>=1.1.0
51
- Requires-Dist: openai-harmony>=0.0.1
52
- Requires-Dist: asyncpg>=0.30.0
53
- Requires-Dist: aiohttp>=3.8.0
54
- Requires-Dist: datasets>=4.0.0
55
- Requires-Dist: transformers>=4.56.1
56
- Requires-Dist: modal>=1.1.4
57
- Provides-Extra: dev
58
- Requires-Dist: build>=1.2.2.post1; extra == "dev"
59
- Requires-Dist: twine>=4.0.0; extra == "dev"
60
- Requires-Dist: keyring>=24.0.0; extra == "dev"
61
- Requires-Dist: pytest>=8.3.3; extra == "dev"
62
- Requires-Dist: pytest-asyncio>=0.24.0; extra == "dev"
63
- Requires-Dist: pytest-cov>=4.1.0; extra == "dev"
64
- Requires-Dist: pyright>=1.1.350; extra == "dev"
65
- Requires-Dist: coverage[toml]>=7.3.0; extra == "dev"
66
- Requires-Dist: ruff>=0.1.0; extra == "dev"
67
- Provides-Extra: research
68
- Requires-Dist: crafter>=1.8.3; extra == "research"
69
- Requires-Dist: datasets>=4.0.0; extra == "research"
70
- Provides-Extra: all
71
- Requires-Dist: crafter>=1.8.3; extra == "all"
72
- Requires-Dist: datasets>=4.0.0; extra == "all"
73
- Dynamic: license-file
74
-
75
- # Synth-AI
76
-
77
- [![Python](https://img.shields.io/badge/python-3.11+-blue)](https://www.python.org/)
78
- [![License](https://img.shields.io/badge/license-MIT-green)](LICENSE)
79
- [![PyPI](https://img.shields.io/badge/PyPI-0.2.4.dev9-orange)](https://pypi.org/project/synth-ai/)
80
- ![Coverage](https://img.shields.io/badge/coverage-0.0%25-red)
81
- ![Tests](https://img.shields.io/badge/tests-17%2F17%20passing-brightgreen)
82
-
83
- Docs: [Synth‑AI Documentation](https://docs.usesynth.ai/synth-ai/introduction)
84
-
85
- Fast and effective reinforcement learning for agents, via an API
86
-
87
- ## Highlights
88
-
89
- - Easily scale gpu topologies - train on 3 a10gs or 8 H100s (multi-node available upon request)
90
- - Requires only a thin fastapi wrapper to integrate with existing agent software.
91
- - Supports the best OSS models like Qwen3. (gpt-oss available upon request, GA soon)
92
- - Own your trained models
93
-
94
- ## Getting Started
95
-
96
- synth-ai comes with a built-in RL example tailored for training Qwen/Qwen3-0.6B to succeed at Math.
97
-
98
- Please create an account at [Synth](https://usesynth.ai) and [Modal](https://modal.com) for the Math hello‑world test run. Then run:
99
-
100
- ```bash
101
- uvx synth-ai demo
102
- uvx synth-ai setup
103
- uvx synth-ai deploy
104
- uvx synth-ai run
105
- ```
106
-
107
- To walk through kicking off your first RL run, see the [Synth‑AI Documentation](https://docs.usesynth.ai/synth-ai/introduction).
108
-
109
- ### What `setup` does now
110
-
111
- When you run `uvx synth-ai setup` (or the legacy `uvx synth-ai rl_demo setup`), the SDK opens your browser to the Synth dashboard for a one‑time pairing (handshake) with your signed‑in session. The SDK will automatically:
112
-
113
- - Detect your current user and organization
114
- - Ensure both API keys exist for that user+org
115
- - Write the keys to your project’s `.env` file as `SYNTH_API_KEY` and `ENVIRONMENT_API_KEY`
116
-
117
- No keys are printed or requested interactively. You’ll see a confirmation like:
118
-
119
- ```
120
- Connecting SDK to your browser session…
121
- Connected to Acme Labs via browser.
122
- ```
123
-
124
- If your browser isn’t already signed in, sign in when prompted and the pairing completes automatically. The dashboard’s welcome modal will reflect a successful pairing.
125
-
126
- Environment variables:
127
-
128
- - `SYNTH_CANONICAL_ORIGIN` (optional): override the dashboard base URL the SDK uses for the handshake (defaults to `http://localhost:3000`).
129
- - Keys are stored only in your project’s `.env` file, not exported to your shell.