auto-coder 0.1.255__py3-none-any.whl → 0.1.256__py3-none-any.whl

{auto_coder-0.1.255.dist-info → auto_coder-0.1.256.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: auto-coder
-Version: 0.1.255
+Version: 0.1.256
 Summary: AutoCoder: AutoCoder
 Author: allwefantasy
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
@@ -26,7 +26,7 @@ Requires-Dist: tabulate
 Requires-Dist: jupyter-client
 Requires-Dist: prompt-toolkit
 Requires-Dist: tokenizers
-Requires-Dist: byzerllm[saas] >=0.1.163
+Requires-Dist: byzerllm[saas] >=0.1.164
 Requires-Dist: patch
 Requires-Dist: diff-match-patch
 Requires-Dist: GitPython

{auto_coder-0.1.255.dist-info → auto_coder-0.1.256.dist-info}/RECORD

@@ -1,5 +1,5 @@
 autocoder/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-autocoder/auto_coder.py,sha256=uzNAtguu7O2m7vve1H2FQnWCmVuqT07RITxS5jgWz4U,64774
+autocoder/auto_coder.py,sha256=obRx5lFu8P4M5mmLA717lwMso7Ei3Kx3i9kG7I_xrRY,65654
 autocoder/auto_coder_lang.py,sha256=Rtupq6N3_HT7JRhDKdgCBcwRaiAnyCOR_Gsp4jUomrI,3229
 autocoder/auto_coder_rag.py,sha256=DDAmqw36CO6phtdQuN8LYIbIR3YGdoZw5-pG0LjVxMc,29063
 autocoder/auto_coder_rag_client_mcp.py,sha256=WV7j5JUiQge0x4-B7Hp5-pSAFXLbvLpzQMcCovbauIM,6276
@@ -10,8 +10,8 @@ autocoder/chat_auto_coder.py,sha256=a1YEp6OPMzpLbRpr2hrbzF6pRnhVPTVxyZfBiQHFPIw,
 autocoder/chat_auto_coder_lang.py,sha256=1cJrjFGrcOQnuP2LdZpgGDSX4CNaIYI7KZGvEEtj6_Q,18242
 autocoder/command_args.py,sha256=9aYJ-AmPxP1sQh6ciw04FWHjSn31f2W9afXFwo8wgx4,30441
 autocoder/lang.py,sha256=U6AjVV8Rs1uLyjFCZ8sT6WWuNUxMBqkXXIOs4S120uk,14511
-autocoder/models.py,sha256=0f653gjpQN_JO5k7h6wmTF4bVd6CW3fpQOyHIZ3ZUv4,7558
-autocoder/version.py,sha256=vxd_TSbCxiXnBHQnfprB361BOJEhjQMPeDRseFI3YrY,23
+autocoder/models.py,sha256=xwWPcegwx945g433UZXna-7HBdnHWCq8oEfHm-HKIDQ,8651
+autocoder/version.py,sha256=Mlq4zYTZeRq2mquyM-8m1qr6sjxAHZpSDVyjrKqhayc,23
 autocoder/agent/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 autocoder/agent/auto_demand_organizer.py,sha256=NWSAEsEk94vT3lGjfo25kKLMwYdPcpy9e-i21txPasQ,6942
 autocoder/agent/auto_filegroup.py,sha256=CW7bqp0FW1GIEMnl-blyAc2UGT7O9Mom0q66ITz1ckM,6635
@@ -29,20 +29,20 @@ autocoder/common/__init__.py,sha256=6maackdzrYnUPvpgVPl92JdMOnw7X4n3EnEQA9OnLGE,
 autocoder/common/anything2images.py,sha256=0ILBbWzY02M-CiWB-vzuomb_J1hVdxRcenAfIrAXq9M,25283
 autocoder/common/anything2img.py,sha256=4TREa-sOA-iargieUy7MpyCYVUE-9Mmq0wJtwomPqnE,7662
 autocoder/common/audio.py,sha256=Kn9nWKQddWnUrAz0a_ZUgjcu4VUU_IcZBigT7n3N3qc,7439
-autocoder/common/auto_coder_lang.py,sha256=nF8XrHpSbibk6ro8Oum-0V0FXcDS1lHP_hL-CnPJtT4,16974
+autocoder/common/auto_coder_lang.py,sha256=x9Zjwvu9OZJjTmswwyimlMb1pvngUAF9_3oNQQut2i4,17634
 autocoder/common/buildin_tokenizer.py,sha256=L7d5t39ZFvUd6EoMPXUhYK1toD0FHlRH1jtjKRGokWU,1236
 autocoder/common/chunk_validation.py,sha256=BrR_ZWavW8IANuueEE7hS8NFAwEvm8TX34WnPx_1hs8,3030
 autocoder/common/cleaner.py,sha256=NU72i8C6o9m0vXExab7nao5bstBUsfJFcj11cXa9l4U,1089
 autocoder/common/code_auto_execute.py,sha256=4KXGmiGObr_B1d6tzV9dwS6MifCSc3Gm4j2d6ildBXQ,6867
-autocoder/common/code_auto_generate.py,sha256=N_kENWW4eVPKmihpIQ4Q2ivYkugAuCWIPLDCKiarRWg,10818
-autocoder/common/code_auto_generate_diff.py,sha256=QAm-BVJgGl_wbw0VjY2RZigVCwOBO-l1W-90PrYDdOQ,16741
-autocoder/common/code_auto_generate_editblock.py,sha256=FuVH-sP4_Ys2MJk9SkWgHHLAKp51bKSxTwFzd7voilg,18356
-autocoder/common/code_auto_generate_strict_diff.py,sha256=kieVUx2KuR0BGtQHE0Ll7i8e_3Lb4xfFXHIXDttQACE,15274
+autocoder/common/code_auto_generate.py,sha256=E8r3VI88hPBPhU5t56qnmeL_fWtCWX1CJvaOachGa2Y,12014
+autocoder/common/code_auto_generate_diff.py,sha256=dmMgN1yIOjJfiYFnzXZuktVFdj4_XR_Tavwx_ysm53U,17846
+autocoder/common/code_auto_generate_editblock.py,sha256=NI_dFwy1VhvdjvARb04-B1AGfgW9z4P1BfWMm-blnaU,19447
+autocoder/common/code_auto_generate_strict_diff.py,sha256=uf5P5B8ly0MP3jCK2PaYJiPLktd1cRRPouwkkaf-DfY,16457
 autocoder/common/code_auto_merge.py,sha256=-ksBjj4ZVcbY_tVH4JLXAMSRtsgaSxrSZ5-MOl9cAgE,7354
 autocoder/common/code_auto_merge_diff.py,sha256=qpEuHJEgX6sWK7EDFEKqcYkyI28wOyM4pytyl8BLohY,15350
 autocoder/common/code_auto_merge_editblock.py,sha256=sxgYMLMACRwJvw-bABkdDHezPelsDFrOCpGuhtT5Dzs,17504
 autocoder/common/code_auto_merge_strict_diff.py,sha256=P0nKNkBrFMybTSZ7kOdA_JixoVmLCZIhAP5q7ILJ9j0,9538
-autocoder/common/code_modification_ranker.py,sha256=qfadP9P-iiidCG2A_MjAf3Ca8cMz7YlnN08D_kH6uFc,6447
+autocoder/common/code_modification_ranker.py,sha256=oG9rCekGsYwE9gNdkIKQ6uKt6uaXpwrC17-FV5Wo-fQ,8187
 autocoder/common/command_completer.py,sha256=IShrZJSpR-Q_MCj_aCVdVyscLYDKj5ZQK357QBcQ_oQ,9420
 autocoder/common/command_generator.py,sha256=-hmbD_AnCa5HxL4BznuEfYAf_l8AxU5fAG5F0sM_fuE,2116
 autocoder/common/command_templates.py,sha256=mnB3n8i0yjH1mqzyClEg8Wpr9VbZV44kxky66Zu6OJY,8557
@@ -76,10 +76,10 @@ autocoder/db/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 autocoder/db/store.py,sha256=tFT66bP2ZKIqZip-uhLkHRSLaaOAUUDZfozJwcqix3c,1908
 autocoder/dispacher/__init__.py,sha256=YoA64dIxnx4jcE1pwSfg81sjkQtjDkhddkfac1-cMWo,1230
 autocoder/dispacher/actions/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-autocoder/dispacher/actions/action.py,sha256=AmckPmTHFKRPTWi4lcw0kAnFw8dyQ9KKoN40JGMuP2s,21971
+autocoder/dispacher/actions/action.py,sha256=sfh3pCasy622Jm0_AIKU7xtR-tqY2tS2_9YJsEd0FJY,22753
 autocoder/dispacher/actions/copilot.py,sha256=iMh4ckj9hO5Q-iemF3CStXd7DatWai7Eci5zOlKxK9c,13072
 autocoder/dispacher/actions/plugins/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-autocoder/dispacher/actions/plugins/action_regex_project.py,sha256=ckTbisMlvwMNHQbrt5WB7pBvf2XAhYQYGH8uyYrvGXU,6060
+autocoder/dispacher/actions/plugins/action_regex_project.py,sha256=22EZL3mLFxgsEZ8ymPCGvaHCJFnrW6C_prp1ykYCuEY,6335
 autocoder/dispacher/actions/plugins/action_translate.py,sha256=nVAtRSQpdGNmZxg1R_9zXG3AuTv3CHf2v7ODgj8u65c,7727
 autocoder/index/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 autocoder/index/entry.py,sha256=1KIGPCtxQN0OdErAco9OmGTd5hB8WJTpWGrxsGLsTcE,12634
@@ -89,8 +89,8 @@ autocoder/index/symbols_utils.py,sha256=CjcjUVajmJZB75Ty3a7kMv1BZphrm-tIBAdOJv6u
 autocoder/index/types.py,sha256=a2s_KV5FJlq7jqA2ELSo9E1sjuLwDB-JJYMhSpzBAhU,596
 autocoder/index/filter/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 autocoder/index/filter/normal_filter.py,sha256=APu34iSvWhtlLtWgkj8N3Vo4oW1TegtZQq2bwDX_cs4,8031
-autocoder/index/filter/quick_filter.py,sha256=Dsm23Z_RrJ_UwCypGUPN1BlKUMibae_9_D8jWD1UDFw,10518
-autocoder/pyproject/__init__.py,sha256=dQ2_7YZ7guybT9BhfxSGn43eLQJGQN2zgeKa6--JlaQ,14403
+autocoder/index/filter/quick_filter.py,sha256=5toipv7XwLsmG_UaqrElpGNjKXq_0bcvFr8W80vT44g,15206
+autocoder/pyproject/__init__.py,sha256=bRuGxFV4QyE85xVjDzeMFmlLVqGbbcFs09FI15Uss4Q,14423
 autocoder/rag/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 autocoder/rag/api_server.py,sha256=dRbhAZVRAOlZ64Cnxf4_rKb4iJwHnrWS9Zr67IVORw0,7288
 autocoder/rag/doc_filter.py,sha256=ZCixxUXNBbz6UiGbgXvbDWdn5moLac3HnZEphpasTDc,6579
@@ -124,14 +124,14 @@ autocoder/rag/stream_event/event_writer.py,sha256=l7kq_LnDDE8E5dZ-73C7J2MgzSL7Wr
 autocoder/rag/stream_event/types.py,sha256=rtLwOE8rShmi1dJdxyBpAV5ZjLBGG9vptMiSzMxGuIA,318
 autocoder/regex_project/__init__.py,sha256=EBZeCL5ORyD_9_5u_UuG4s7XtpXOu0y1sWDmxWFtufE,6781
 autocoder/regexproject/__init__.py,sha256=cEr-ZOaQjLD5sx7T7F2DhD5ips03HcJ02rded9EpSXc,9693
-autocoder/suffixproject/__init__.py,sha256=VcXjUbGf3uQrpoqVCItDvGG9DoeHJ_qEmghKwrVNw9w,11058
-autocoder/tsproject/__init__.py,sha256=boNuRCHi94xI_y4tvL5LKgSZ4gYxcPqUUQTw9MU_STI,11751
+autocoder/suffixproject/__init__.py,sha256=2dxh8vizDLiSqGpCx-V2VQ6cOfRZGxJMgosjwSlAsQM,11078
+autocoder/tsproject/__init__.py,sha256=zGVSrxFo15Nh8GcSXHwdk4Fr-W7Bdb6IwVUX46PZKy4,11771
 autocoder/utils/__init__.py,sha256=KtcGElFNBgZPF7dEL8zF9JpXkCAjoyDrzaREJBhJrcs,994
 autocoder/utils/_markitdown.py,sha256=RU88qn4eZfYIy0GDrPxlI8oYXIypbi63VRJjdlnE0VU,47431
 autocoder/utils/coder.py,sha256=rK8e0svQBe0NOP26dIGToUXgha_hUDgxlWoC_p_r7oc,5698
 autocoder/utils/conversation_store.py,sha256=sz-hhY7sttPAUOAQU6Pze-5zJc3j0_Emj22dM_0l5ro,1161
 autocoder/utils/llm_client_interceptors.py,sha256=FEHNXoFZlCjAHQcjPRyX8FOMjo6rPXpO2AJ2zn2KTTo,901
-autocoder/utils/llms.py,sha256=RBOSzH6xhedISzmiQMGH_swmVfY-QSTe5Cm1ZZGBhNE,2948
+autocoder/utils/llms.py,sha256=HM5K_v4AcuWo65lgcp66DEqaU9-fjoT7mcI1iv2Fopg,3839
 autocoder/utils/log_capture.py,sha256=I-bsJFLWoGUiX-GKoZsH9kWJCKSV7ZlUnRt7jh-fOL0,1548
 autocoder/utils/multi_turn.py,sha256=unK9OpqVRbK6uIcTKXgggX2wNmyj7s5eyEAQ2xUwHoM,88
 autocoder/utils/operate_config_api.py,sha256=99YAKsuUFLPwrRvj0CJal_bAPgyiXWMma6ZKMU56thw,5790
@@ -146,9 +146,9 @@ autocoder/utils/types.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 autocoder/utils/auto_coder_utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 autocoder/utils/auto_coder_utils/chat_stream_out.py,sha256=xWXqICANbDOovH4wcFW1eSI7lB7TjXbk1mSU4bTKEW4,11434
 autocoder/utils/chat_auto_coder_utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-auto_coder-0.1.255.dist-info/LICENSE,sha256=HrhfyXIkWY2tGFK11kg7vPCqhgh5DcxleloqdhrpyMY,11558
-auto_coder-0.1.255.dist-info/METADATA,sha256=2uB08jgGHyp3_DWMI2_vxoFoptVJO76Va-yek1umBac,2616
-auto_coder-0.1.255.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
-auto_coder-0.1.255.dist-info/entry_points.txt,sha256=0nzHtHH4pNcM7xq4EBA2toS28Qelrvcbrr59GqD_0Ak,350
-auto_coder-0.1.255.dist-info/top_level.txt,sha256=Jqc0_uJSw2GwoFQAa9iJxYns-2mWla-9ok_Y3Gcznjk,10
-auto_coder-0.1.255.dist-info/RECORD,,
+auto_coder-0.1.256.dist-info/LICENSE,sha256=HrhfyXIkWY2tGFK11kg7vPCqhgh5DcxleloqdhrpyMY,11558
+auto_coder-0.1.256.dist-info/METADATA,sha256=NrC0Y2oSS6lDTeKbXKk3F1QnZ7_3Kie7pho_i7JhwJk,2616
+auto_coder-0.1.256.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
+auto_coder-0.1.256.dist-info/entry_points.txt,sha256=0nzHtHH4pNcM7xq4EBA2toS28Qelrvcbrr59GqD_0Ak,350
+auto_coder-0.1.256.dist-info/top_level.txt,sha256=Jqc0_uJSw2GwoFQAa9iJxYns-2mWla-9ok_Y3Gcznjk,10
+auto_coder-0.1.256.dist-info/RECORD,,

autocoder/auto_coder.py

@@ -1389,11 +1389,25 @@ def main(input_args: Optional[List[str]] = None):
                 elapsed_time = time.time() - start_time
                 printer = Printer()
                 speed = last_meta.generated_tokens_count / elapsed_time
+                
+                # Get model info for pricing
+                from autocoder.utils import llms as llm_utils
+                model_info = llm_utils.get_model_info(model_name, args.product_mode) or {}
+                input_price = model_info.get("input_price", 0.0) if model_info else 0.0
+                output_price = model_info.get("output_price", 0.0) if model_info else 0.0
+                
+                # Calculate costs
+                input_cost = (last_meta.input_tokens_count * input_price) / 1000000  # Convert to millions
+                output_cost = (last_meta.generated_tokens_count * output_price) / 1000000  # Convert to millions
+                
                 printer.print_in_terminal("stream_out_stats", 
+                                    model_name=model_name,
                                     elapsed_time=elapsed_time,
                                     first_token_time=last_meta.first_token_time,
                                     input_tokens=last_meta.input_tokens_count,
                                     output_tokens=last_meta.generated_tokens_count,
+                                    input_cost=round(input_cost, 4),
+                                    output_cost=round(output_cost, 4),
                                     speed=round(speed, 2))
             
             chat_history["ask_conversation"].append(

autocoder/common/auto_coder_lang.py

@@ -57,7 +57,7 @@ MESSAGES = {
             "Paste the answer to the input box below, use '/break' to exit, '/clear' to clear the screen, '/eof' to submit."
         ),
         "code_generation_start": "Auto generate the code...",
-        "code_generation_complete": "Code generation completed in {{ duration }} seconds, input_tokens_count: {{ input_tokens }}, generated_tokens_count: {{ output_tokens }}, speed: {{ speed }} tokens/s",
+        "code_generation_complete": "{{ model_names}} Code generation completed in {{ duration }} seconds, input_tokens_count: {{ input_tokens }}, generated_tokens_count: {{ output_tokens }}, input_cost: {{ input_cost }}, output_cost: {{ output_cost }}, speed: {{ speed }} tokens/s",
         "code_merge_start": "Auto merge the code...",
         "code_execution_warning": "Content(send to model) is {{ content_length }} tokens (you may collect too much files), which is larger than the maximum input length {{ max_length }}",
         "quick_filter_start": "{{ model_name }} Starting filter context(quick_filter)...",
@@ -75,12 +75,12 @@ MESSAGES = {
         "ranking_start": "Start ranking {{ count }} candidates using model {{ model_name }}",
         "ranking_failed_request": "Ranking request failed: {{ error }}",
         "ranking_all_failed": "All ranking requests failed",
-        "ranking_complete": "Ranking completed in {{ elapsed }}s, total voters: {{ total_tasks }}, best candidate index: {{ best_candidate }}, scores: {{ scores }}, input_tokens: {{ input_tokens }}, output_tokens: {{ output_tokens }}",
+        "ranking_complete": "{{ model_names }} Ranking completed in {{ elapsed }}s, total voters: {{ total_tasks }}, best candidate index: {{ best_candidate }}, scores: {{ scores }}, input_tokens: {{ input_tokens }}, output_tokens: {{ output_tokens }}, input_cost: {{ input_cost }}, output_cost: {{ output_cost }}",
         "ranking_process_failed": "Ranking process failed: {{ error }}",
         "ranking_failed": "Ranking failed in {{ elapsed }}s, using original order",
         "begin_index_source_code": "🚀 Begin to index source code in {{ source_dir }}",
-        "stream_out_stats": "Elapsed time {{ elapsed_time }} seconds, first token time: {{ first_token_time }} seconds, input tokens: {{ input_tokens }}, output tokens: {{ output_tokens }}, speed: {{ speed }} tokens/s",
-        "quick_filter_stats": "快速过滤器完成，耗时 {{ elapsed_time }} 秒，输入token数: {{ input_tokens }}, 输出token数: {{ output_tokens }}",
+        "stream_out_stats": "Model: {{ model_name }}, Total time: {{ elapsed_time }} seconds, First token time: {{ first_token_time }} seconds, Speed: {{ speed }} tokens/s, Input tokens: {{ input_tokens }}, Output tokens: {{ output_tokens }}, Input cost: {{ input_cost }}, Output cost: {{ output_cost }}",
+        "quick_filter_stats": "{{ model_names }} 快速过滤器完成，耗时 {{ elapsed_time }} 秒，输入token数: {{ input_tokens }}, 输出token数: {{ output_tokens }}, 输入成本: {{ input_cost }}, 输出成本: {{ output_cost }}",
         "upsert_file": "✅ Updated file: {{ file_path }}",  
         "unmerged_blocks_title": "Unmerged Blocks",
         "quick_filter_title": "{{ model_name }} is analyzing how to filter context...",
@@ -150,7 +150,7 @@ MESSAGES = {
             "将获得答案黏贴到下面的输入框，换行后，使用 '/break' 退出，'/clear' 清屏，'/eof' 提交。"
         ),
         "code_generation_start": "正在自动生成代码...",
-        "code_generation_complete": "代码生成完成，耗时 {{ duration }} 秒，输入token数: {{ input_tokens }}, 输出token数: {{ output_tokens }}, 速度: {{ speed }} tokens/秒",
+        "code_generation_complete": "{{ model_names}} 代码生成完成，耗时 {{ duration }} 秒，输入token数: {{ input_tokens }}, 输出token数: {{ output_tokens }}, 输入成本: {{ input_cost }}, 输出成本: {{ output_cost }}, 速度: {{ speed }} tokens/秒",
         "code_merge_start": "正在自动合并代码...",
         "code_execution_warning": "发送给模型的内容长度为 {{ content_length }} tokens（您可能收集了太多文件），超过了最大输入长度 {{ max_length }}",
         "quick_filter_start": "{{ model_name }} 开始查找上下文(quick_filter)...",
@@ -179,11 +179,11 @@ MESSAGES = {
         "ranking_start": "开始对 {{ count }} 个候选项进行排序,使用模型 {{ model_name }} 打分",
         "ranking_failed_request": "排序请求失败: {{ error }}",
         "ranking_all_failed": "所有排序请求都失败",
-        "ranking_complete": "排序完成，耗时 {{ elapsed }} 秒，总投票数: {{ total_tasks }}，最佳候选索引: {{ best_candidate }}，得分: {{ scores }}，输入token数: {{ input_tokens }}，输出token数: {{ output_tokens }}",
+        "ranking_complete": "{{ model_names }} 排序完成，耗时 {{ elapsed }} 秒，总投票数: {{ total_tasks }}，最佳候选索引: {{ best_candidate }}，得分: {{ scores }}，输入token数: {{ input_tokens }}，输出token数: {{ output_tokens }}  输入成本: {{ input_cost }}, 输出成本: {{ output_cost }}",
         "ranking_process_failed": "排序过程失败: {{ error }}",
         "ranking_failed": "排序失败，耗时 {{ elapsed }} 秒，使用原始顺序",
-        "stream_out_stats": "总耗时 {{ elapsed_time }} 秒，首token时间: {{ first_token_time }} 秒，输入token数: {{ input_tokens }}, 输出token数: {{ output_tokens }}, 速度: {{ speed }} tokens/秒",
-        "quick_filter_stats": "Quick filter completed in {{ elapsed_time }} seconds, input tokens: {{ input_tokens }}, output tokens: {{ output_tokens }}",
+        "stream_out_stats": "模型: {{ model_name }},总耗时 {{ elapsed_time }} 秒,首token时间: {{ first_token_time }} 秒, 速度: {{ speed }} tokens/秒, 输入token数: {{ input_tokens }}, 输出token数: {{ output_tokens }}, 输入成本: {{ input_cost }}, 输出成本: {{ output_cost }}",
+        "quick_filter_stats": "{{ model_names }} Quick filter completed in {{ elapsed_time }} seconds, input tokens: {{ input_tokens }}, output tokens: {{ output_tokens }}, input cost: {{ input_cost }}, output cost: {{ output_cost }}",
         "quick_filter_title": "{{ model_name }} 正在分析如何筛选上下文...",
         "quick_filter_failed": "❌ 快速过滤器失败: {{ error }}. ",
         "estimated_chat_input_tokens": "对话输入token预估为: {{ estimated_input_tokens }}",

autocoder/common/code_auto_generate.py

@@ -10,6 +10,7 @@ from autocoder.common.utils_code_auto_generate import chat_with_continue
 import json
 from autocoder.common.printer import Printer
 from autocoder.rag.token_counter import count_tokens
+from autocoder.utils import llms as llm_utils
 
 
 class CodeAutoGenerate:
@@ -193,6 +194,9 @@ class CodeAutoGenerate:
         results = []
         input_tokens_count = 0
         generated_tokens_count = 0
+        input_tokens_cost = 0
+        generated_tokens_cost = 0
+        model_names = []
 
         printer = Printer()
         estimated_input_tokens = count_tokens(json.dumps(conversations, ensure_ascii=False))
@@ -206,13 +210,27 @@ class CodeAutoGenerate:
                 futures = []
                 for llm in self.llms:
                     for _ in range(self.generate_times_same_model):
-                        futures.append(executor.submit(
-                            chat_with_continue, llm=llm, conversations=conversations, llm_config=llm_config))
+                        
+                        model_names_list = llm_utils.get_llm_names(llm)
+                        model_name = None
+                        if model_names_list:
+                            model_name = model_names_list[0]                                                    
+                        
+                        for _ in range(self.generate_times_same_model):
+                            model_names.append(model_name)
+                            futures.append(executor.submit(
+                                chat_with_continue, llm=llm, conversations=conversations, llm_config=llm_config))
+                            
                 temp_results = [future.result() for future in futures]
                 for result in temp_results:
                     results.append(result.content)
                     input_tokens_count += result.input_tokens_count
                     generated_tokens_count += result.generated_tokens_count
+                    model_info = llm_utils.get_model_info(model_name, self.args.product_mode)
+                    input_cost = model_info.get("input_price", 0) if model_info else 0
+                    output_cost = model_info.get("output_price", 0) if model_info else 0
+                    input_tokens_cost += input_cost * result.input_tokens_count / 1000000
+                    generated_tokens_cost += output_cost * result.generated_tokens_count / 1000000
             
             for result in results:
                 conversations_list.append(
@@ -227,7 +245,9 @@ class CodeAutoGenerate:
         
         statistics = {
             "input_tokens_count": input_tokens_count,
-            "generated_tokens_count": generated_tokens_count
+            "generated_tokens_count": generated_tokens_count,
+            "input_tokens_cost": input_tokens_cost,
+            "generated_tokens_cost": generated_tokens_cost
         }        
     
         if self.args.request_id and not self.args.skip_events:

autocoder/common/code_auto_generate_diff.py

@@ -9,6 +9,7 @@ import json
 from autocoder.common.utils_code_auto_generate import chat_with_continue
 from autocoder.common.printer import Printer
 from autocoder.rag.token_counter import count_tokens
+from autocoder.utils import llms as llm_utils
 
 
 class CodeAutoGenerateDiff:
@@ -341,6 +342,9 @@ class CodeAutoGenerateDiff:
         results = []
         input_tokens_count = 0
         generated_tokens_count = 0
+        input_tokens_cost = 0
+        generated_tokens_cost = 0
+        model_names = []
 
         printer = Printer()
         estimated_input_tokens = count_tokens(json.dumps(conversations, ensure_ascii=False))
@@ -354,13 +358,26 @@ class CodeAutoGenerateDiff:
                 futures = []
                 for llm in self.llms:
                     for _ in range(self.generate_times_same_model):
-                        futures.append(executor.submit(
-                            chat_with_continue, llm=llm, conversations=conversations, llm_config=llm_config))
+                        model_names_list = llm_utils.get_llm_names(llm)
+                        model_name = None
+                        if model_names_list:
+                            model_name = model_names_list[0]
+                        
+                        for _ in range(self.generate_times_same_model):
+                            model_names.append(model_name)
+                            futures.append(executor.submit(
+                                chat_with_continue, llm=llm, conversations=conversations, llm_config=llm_config))
+                
                 temp_results = [future.result() for future in futures]
                 for result in temp_results:
                     results.append(result.content)
                     input_tokens_count += result.input_tokens_count
                     generated_tokens_count += result.generated_tokens_count
+                    model_info = llm_utils.get_model_info(model_name, self.args.product_mode)
+                    input_cost = model_info.get("input_price",0) if model_info else 0
+                    output_cost = model_info.get("output_price",0) if model_info else 0
+                    input_tokens_cost += input_cost * result.input_tokens_count / 1000000
+                    generated_tokens_cost += output_cost * result.generated_tokens_count / 1000000
             
             for result in results:
                 conversations_list.append(
@@ -376,7 +393,9 @@ class CodeAutoGenerateDiff:
         
         statistics = {
             "input_tokens_count": input_tokens_count,
-            "generated_tokens_count": generated_tokens_count
+            "generated_tokens_count": generated_tokens_count,
+            "input_tokens_cost": input_tokens_cost,
+            "generated_tokens_cost": generated_tokens_cost
         }        
     
         if self.args.request_id and not self.args.skip_events:

autocoder/common/code_auto_generate_editblock.py

@@ -13,6 +13,7 @@ from concurrent.futures import ThreadPoolExecutor
 from autocoder.common.utils_code_auto_generate import chat_with_continue
 from autocoder.common.printer import Printer
 from autocoder.rag.token_counter import count_tokens
+from autocoder.utils import llms as llm_utils
 
 
 class CodeAutoGenerateEditBlock:
@@ -424,6 +425,11 @@ class CodeAutoGenerateEditBlock:
         input_tokens_count = 0
         generated_tokens_count = 0
 
+        input_tokens_cost = 0
+        generated_tokens_cost = 0
+
+        model_names = []
+
         printer = Printer()
         estimated_input_tokens = count_tokens(
             json.dumps(conversations, ensure_ascii=False))
@@ -437,14 +443,28 @@ class CodeAutoGenerateEditBlock:
             with ThreadPoolExecutor(max_workers=len(self.llms) * self.generate_times_same_model) as executor:
                 futures = []
                 for llm in self.llms:
+                    
+                    model_names_list = llm_utils.get_llm_names(llm) 
+                    model_name = None
+                    if model_names_list:
+                        model_name = model_names_list[0]                    
+
                     for _ in range(self.generate_times_same_model):
+                        model_names.append(model_name)
                         futures.append(executor.submit(
                             chat_with_continue, llm=llm, conversations=conversations, llm_config=llm_config))
+                
                 temp_results = [future.result() for future in futures]
-                for result in temp_results:
+                
+                for result,model_name in zip(temp_results,model_names):
                     results.append(result.content)
                     input_tokens_count += result.input_tokens_count
                     generated_tokens_count += result.generated_tokens_count
+                    model_info = llm_utils.get_model_info(model_name,self.args.product_mode)                    
+                    input_cost = model_info.get("input_price", 0) if model_info else 0
+                    output_cost = model_info.get("output_price", 0) if model_info else 0
+                    input_tokens_cost += input_cost * result.input_tokens_count / 1000000
+                    generated_tokens_cost += output_cost * result.generated_tokens_count / 1000000
 
             for result in results:
                 conversations_list.append(
@@ -461,7 +481,9 @@ class CodeAutoGenerateEditBlock:
 
         statistics = {
             "input_tokens_count": input_tokens_count,
-            "generated_tokens_count": generated_tokens_count
+            "generated_tokens_count": generated_tokens_count,
+            "input_tokens_cost": input_tokens_cost,
+            "generated_tokens_cost": generated_tokens_cost
         }
 
         if self.args.request_id and not self.args.skip_events:

autocoder/common/code_auto_generate_strict_diff.py

@@ -9,6 +9,7 @@ import json
 from autocoder.common.utils_code_auto_generate import chat_with_continue
 from autocoder.common.printer import Printer
 from autocoder.rag.token_counter import count_tokens
+from autocoder.utils import llms as llm_utils
 
 class CodeAutoGenerateStrictDiff:
     def __init__(
@@ -311,6 +312,9 @@ class CodeAutoGenerateStrictDiff:
         results = []
         input_tokens_count = 0
         generated_tokens_count = 0
+        input_tokens_cost = 0
+        generated_tokens_cost = 0
+        model_names = []
 
         printer = Printer()
         estimated_input_tokens = count_tokens(json.dumps(conversations, ensure_ascii=False))
@@ -324,14 +328,27 @@ class CodeAutoGenerateStrictDiff:
                 futures = []
                 for llm in self.llms:
                     for _ in range(self.generate_times_same_model):
-                        futures.append(executor.submit(
-                            chat_with_continue, llm=llm, conversations=conversations, llm_config=llm_config))
+                        
+                        model_names_list = llm_utils.get_llm_names(llm)
+                        model_name = None
+                        if model_names_list:
+                            model_name = model_names_list[0]                                                    
+                        
+                        for _ in range(self.generate_times_same_model):
+                            model_names.append(model_name)
+                            futures.append(executor.submit(
+                                chat_with_continue, llm=llm, conversations=conversations, llm_config=llm_config))
+                            
                 temp_results = [future.result() for future in futures]
                 for result in temp_results:
                     results.append(result.content)
                     input_tokens_count += result.input_tokens_count
                     generated_tokens_count += result.generated_tokens_count
-            
+                    model_info = llm_utils.get_model_info(model_name, self.args.product_mode)
+                    input_cost = model_info.get("input_price", 0) if model_info else 0
+                    output_cost = model_info.get("output_price", 0) if model_info else 0
+                    input_tokens_cost += input_cost * result.input_tokens_count / 1000000
+                    generated_tokens_cost += output_cost * result.generated_tokens_count / 1000000
             for result in results:
                 conversations_list.append(
                     conversations + [{"role": "assistant", "content": result}])
@@ -345,7 +362,9 @@ class CodeAutoGenerateStrictDiff:
         
         statistics = {
             "input_tokens_count": input_tokens_count,
-            "generated_tokens_count": generated_tokens_count
+            "generated_tokens_count": generated_tokens_count,
+            "input_tokens_cost": input_tokens_cost,
+            "generated_tokens_cost": generated_tokens_cost
         }        
     
         if self.args.request_id and not self.args.skip_events:

autocoder/common/code_modification_ranker.py

@@ -8,8 +8,8 @@ from concurrent.futures import ThreadPoolExecutor, as_completed
 import traceback
 from autocoder.common.utils_code_auto_generate import chat_with_continue
 from byzerllm.utils.str2model import to_model
+from autocoder.utils.llms import get_llm_names, get_model_info
 
-from autocoder.utils.llms import get_llm_names
 class RankResult(BaseModel):
     rank_result: List[int]
 
@@ -97,13 +97,42 @@ class CodeModificationRanker:
 
                 # Collect all results
                 results = []
-                for future in as_completed(futures):
+                # 获取模型名称列表
+                model_names = []
+                for llm in self.llms:
+                    # 获取当前llm实例对应的模型名称
+                    names = get_llm_names(llm)  
+                    model_names.extend(names)
+
+                # 获取模型价格信息
+                model_info_map = {}
+                for name in model_names:
+                    # 第二个参数是产品模式,从args中获取
+                    info = get_model_info(name, self.args.product_mode)  
+                    if info:
+                        model_info_map[name] = {
+                            "input_cost": info.get("input_price", 0.0),  # 每百万tokens成本
+                            "output_cost": info.get("output_price", 0.0) # 每百万tokens成本 
+                        }
+
+                # 计算总成本
+                total_input_cost = 0.0
+                total_output_cost = 0.0
+
+                for future, model_name in zip(futures, model_names):
                     try:
                         result = future.result()
                         input_tokens_count += result.input_tokens_count
                         generated_tokens_count += result.generated_tokens_count
                         v = to_model(result.content,RankResult)                        
                         results.append(v.rank_result)
+
+                        # 计算成本                        
+                        info = model_info_map.get(model_name, {})
+                        # 计算公式:token数 * 单价 / 1000000
+                        total_input_cost += (result.input_tokens_count * info.get("input_cost", 0.0)) / 1000000
+                        total_output_cost += (result.generated_tokens_count * info.get("output_cost", 0.0)) / 1000000
+
                     except Exception as e:
                         self.printer.print_in_terminal(
                             "ranking_failed_request", style="yellow", error=str(e))                        
@@ -113,6 +142,10 @@ class CodeModificationRanker:
                     raise Exception(
                         self.printer.get_message_from_key("ranking_all_failed"))
 
+                # 四舍五入到4位小数
+                total_input_cost = round(total_input_cost, 4)
+                total_output_cost = round(total_output_cost, 4)
+
                 # Calculate scores for each candidate
                 candidate_scores = defaultdict(float)
                 for rank_result in results:
@@ -137,7 +170,10 @@ class CodeModificationRanker:
                     best_candidate=sorted_candidates[0],
                     scores=score_details,
                     input_tokens=input_tokens_count,
-                    output_tokens=generated_tokens_count
+                    output_tokens=generated_tokens_count,
+                    input_cost=total_input_cost,
+                    output_cost=total_output_cost,
+                    model_names=", ".join(model_names)
                 )
 
                 rerank_contents = [generate_result.contents[i]

autocoder/dispacher/actions/action.py

@@ -125,13 +125,17 @@ class ActionTSProject(BaseAction):
                     query=args.query, source_content=content
                 )
             elapsed_time = time.time() - start_time
-            speed = generate_result.metadata.get('generated_tokens_count', 0) / elapsed_time if elapsed_time > 0 else 0            
-            model_names = ",".join(get_llm_names(self.llm))
+            speed = generate_result.metadata.get('generated_tokens_count', 0) / elapsed_time if elapsed_time > 0 else 0
+            input_tokens_cost = generate_result.metadata.get('input_tokens_cost', 0)
+            generated_tokens_cost = generate_result.metadata.get('generated_tokens_cost', 0)
+            model_names = ",".join(get_llm_names(generate.llms))
             self.printer.print_in_terminal(
                 "code_generation_complete",
                 duration=elapsed_time,
                 input_tokens=generate_result.metadata.get('input_tokens_count', 0),
                 output_tokens=generate_result.metadata.get('generated_tokens_count', 0),
+                input_cost=input_tokens_cost,
+                output_cost=generated_tokens_cost,
                 speed=round(speed, 2),
                 model_names=model_names
             )
@@ -221,12 +225,16 @@ class ActionPyScriptProject(BaseAction):
 
             elapsed_time = time.time() - start_time
             speed = generate_result.metadata.get('generated_tokens_count', 0) / elapsed_time if elapsed_time > 0 else 0
-            model_names = ",".join(get_llm_names(self.llm))
+            model_names = ",".join(get_llm_names(generate.llms))
+            input_tokens_cost = generate_result.metadata.get('input_tokens_cost', 0)
+            generated_tokens_cost = generate_result.metadata.get('generated_tokens_cost', 0)
             self.printer.print_in_terminal(
                 "code_generation_complete",
                 duration=elapsed_time,
                 input_tokens=generate_result.metadata.get('input_tokens_count', 0),
                 output_tokens=generate_result.metadata.get('generated_tokens_count', 0),
+                input_cost=input_tokens_cost,
+                output_cost=generated_tokens_cost,
                 speed=round(speed, 2),
                 model_names=model_names
             )
@@ -264,13 +272,7 @@ class ActionPyScriptProject(BaseAction):
                     model=self.llm.default_model_name,
                 )
 
-            end_time = time.time()
-            self.printer.print_in_terminal(
-                "code_generation_complete",
-                duration=end_time - start_time,
-                input_tokens=generate_result.metadata.get('input_tokens_count', 0),
-                output_tokens=generate_result.metadata.get('generated_tokens_count', 0)
-            )
+            end_time = time.time()            
             with open(self.args.target_file, "w") as file:
                 file.write(content)
 
@@ -348,12 +350,16 @@ class ActionPyProject(BaseAction):
                 )
             elapsed_time = time.time() - start_time
             speed = generate_result.metadata.get('generated_tokens_count', 0) / elapsed_time if elapsed_time > 0 else 0
-            model_names = ",".join(get_llm_names(self.llm))
+            model_names = ",".join(get_llm_names(generate.llms))
+            input_tokens_cost = generate_result.metadata.get('input_tokens_cost', 0)
+            generated_tokens_cost = generate_result.metadata.get('generated_tokens_cost', 0)
             self.printer.print_in_terminal(
                 "code_generation_complete",
                 duration=elapsed_time,
                 input_tokens=generate_result.metadata.get('input_tokens_count', 0),
                 output_tokens=generate_result.metadata.get('generated_tokens_count', 0),
+                input_cost=input_tokens_cost,
+                output_cost=generated_tokens_cost,
                 speed=round(speed, 2),
                 model_names=model_names
             )
@@ -458,12 +464,16 @@ class ActionSuffixProject(BaseAction):
               
         elapsed_time = time.time() - start_time
         speed = generate_result.metadata.get('generated_tokens_count', 0) / elapsed_time if elapsed_time > 0 else 0
-        model_names = ",".join(get_llm_names(self.llm))
+        model_names = ",".join(get_llm_names(generate.llms))
+        input_tokens_cost = generate_result.metadata.get('input_tokens_cost', 0)
+        generated_tokens_cost = generate_result.metadata.get('generated_tokens_cost', 0)
         self.printer.print_in_terminal(
             "code_generation_complete",
             duration=elapsed_time,
             input_tokens=generate_result.metadata.get('input_tokens_count', 0),
             output_tokens=generate_result.metadata.get('generated_tokens_count', 0),
+            input_cost=input_tokens_cost,
+            output_cost=generated_tokens_cost,
             speed=round(speed, 2),
             model_names=model_names
         )

autocoder/dispacher/actions/plugins/action_regex_project.py

@@ -88,11 +88,15 @@ class ActionRegexProject:
             elapsed_time = time.time() - start_time
             speed = generate_result.metadata.get('generated_tokens_count', 0) / elapsed_time if elapsed_time > 0 else 0
             model_names = ",".join(get_llm_names(self.llm))
+            input_tokens_cost = generate_result.metadata.get('input_tokens_cost', 0)
+            generated_tokens_cost = generate_result.metadata.get('generated_tokens_cost', 0)
             self.printer.print_in_terminal(
                 "code_generation_complete",
                 duration=elapsed_time,
                 input_tokens=generate_result.metadata.get('input_tokens_count', 0),
                 output_tokens=generate_result.metadata.get('generated_tokens_count', 0),
+                input_cost=input_tokens_cost,
+                output_cost=generated_tokens_cost,
                 speed=round(speed, 2),
                 model_names=model_names
             )

autocoder/index/filter/quick_filter.py

@@ -4,21 +4,21 @@ from autocoder.utils.auto_coder_utils.chat_stream_out import stream_out
 from autocoder.common.utils_code_auto_generate import stream_chat_with_continue
 from byzerllm.utils.str2model import to_model
 from autocoder.index.types import IndexItem
-from autocoder.common import AutoCoderArgs,SourceCode
+from autocoder.common import AutoCoderArgs, SourceCode
 import byzerllm
 import time
 from autocoder.index.index import IndexManager
 from autocoder.index.types import (
     IndexItem,
-    TargetFile,    
+    TargetFile,
     FileNumberList
 )
 from autocoder.rag.token_counter import count_tokens
 from autocoder.common.printer import Printer
 from concurrent.futures import ThreadPoolExecutor
-import threading
+from byzerllm import MetaHolder
 
-from autocoder.utils.llms import get_llm_names
+from autocoder.utils.llms import get_llm_names, get_model_info
 
 
 def get_file_path(file_path):
@@ -32,8 +32,9 @@ class QuickFilterResult(BaseModel):
     has_error: bool
     error_message: Optional[str] = None
 
+
 class QuickFilter():
-    def __init__(self, index_manager: IndexManager,stats:Dict[str,Any],sources:List[SourceCode]):
+    def __init__(self, index_manager: IndexManager, stats: Dict[str, Any], sources: List[SourceCode]):
         self.index_manager = index_manager
         self.args = index_manager.args
         self.stats = stats
@@ -41,72 +42,142 @@ class QuickFilter():
         self.printer = Printer()
         self.max_tokens = self.args.index_filter_model_max_input_length
 
-
     def big_filter(self, index_items: List[IndexItem],) -> QuickFilterResult:
         chunks = []
         current_chunk = []
-                
+
         # 将 index_items 切分成多个 chunks,第一个chunk尽可能接近max_tokens
         for item in index_items:
             # 使用 quick_filter_files.prompt 生成文本再统计
             temp_chunk = current_chunk + [item]
-            prompt_text = self.quick_filter_files.prompt(temp_chunk, self.args.query)
-            temp_size = count_tokens(prompt_text)            
+            prompt_text = self.quick_filter_files.prompt(
+                temp_chunk, self.args.query)
+            temp_size = count_tokens(prompt_text)
             # 如果当前chunk为空,或者添加item后不超过max_tokens,就添加到当前chunk
             if not current_chunk or temp_size <= self.max_tokens:
-                current_chunk.append(item)                
+                current_chunk.append(item)
             else:
                 # 当前chunk已满,创建新chunk
                 chunks.append(current_chunk)
-                current_chunk = [item]                
-            
+                current_chunk = [item]
+
         if current_chunk:
             chunks.append(current_chunk)
-        
-        tokens_len = count_tokens(self.quick_filter_files.prompt(index_items, self.args.query))
+
+        tokens_len = count_tokens(
+            self.quick_filter_files.prompt(index_items, self.args.query))
         self.printer.print_in_terminal(
-                "quick_filter_too_long",
-                style="yellow",
-                tokens_len=tokens_len,
-                max_tokens=self.max_tokens,
-                split_size=len(chunks)
-            )    
+            "quick_filter_too_long",
+            style="yellow",
+            tokens_len=tokens_len,
+            max_tokens=self.max_tokens,
+            split_size=len(chunks)
+        )
 
         def process_chunk(chunk_index: int, chunk: List[IndexItem]) -> QuickFilterResult:
             try:
-                model_name = ",".join(get_llm_names(self.index_manager.index_filter_llm))
+                # 获取模型名称列表
+                model_names = get_llm_names(
+                    self.index_manager.index_filter_llm)
+                model_name = ",".join(model_names)
                 files: Dict[str, TargetFile] = {}
-                
+
+                # 获取模型价格信息
+                model_info_map = {}
+                for name in model_names:
+                    # 第二个参数是产品模式,从args中获取
+                    info = get_model_info(name, self.args.product_mode)
+                    if info:
+                        model_info_map[name] = {
+                            # 每百万tokens成本
+                            "input_price": info.get("input_price", 0.0),
+                            # 每百万tokens成本
+                            "output_price": info.get("output_price", 0.0)
+                        }
+
                 if chunk_index == 0:
                     # 第一个chunk使用流式输出
                     stream_generator = stream_chat_with_continue(
                         self.index_manager.index_filter_llm,
-                        [{"role": "user", "content": self.quick_filter_files.prompt(chunk, self.args.query)}],
+                        [{"role": "user", "content": self.quick_filter_files.prompt(
+                            chunk, self.args.query)}],
                         {}
                     )
-                    full_response, _ = stream_out(
+                    full_response, last_meta = stream_out(
                         stream_generator,
                         model_name=model_name,
-                        title=self.printer.get_message_from_key_with_format("quick_filter_title", model_name=model_name),
+                        title=self.printer.get_message_from_key_with_format(
+                            "quick_filter_title", model_name=model_name),
                         args=self.args
                     )
                     file_number_list = to_model(full_response, FileNumberList)
+
+                    # 计算总成本
+                    total_input_cost = 0.0
+                    total_output_cost = 0.0
+
+                    for name in model_names:
+                        info = model_info_map.get(name, {})
+                        # 计算公式:token数 * 单价 / 1000000
+                        total_input_cost += (last_meta.input_tokens_count *
+                                             info.get("input_price", 0.0)) / 1000000
+                        total_output_cost += (last_meta.generated_tokens_count *
+                                              info.get("output_price", 0.0)) / 1000000
+
+                    # 四舍五入到4位小数
+                    total_input_cost = round(total_input_cost, 4)
+                    total_output_cost = round(total_output_cost, 4)
+
+                    # 打印 token 统计信息和成本
+                    self.printer.print_in_terminal(
+                        "quick_filter_stats",
+                        style="blue",
+                        input_tokens=last_meta.input_tokens_count,
+                        output_tokens=last_meta.generated_tokens_count,
+                        input_cost=total_input_cost,
+                        output_cost=total_output_cost,
+                        model_names=model_name
+                    )
                 else:
                     # 其他chunks直接使用with_llm
-                    file_number_list = self.quick_filter_files.with_llm(self.index_manager.index_filter_llm).with_return_type(FileNumberList).run(chunk, self.args.query)
-                
+                    meta_holder = MetaHolder()
+                    start_time = time.monotonic()
+                    file_number_list = self.quick_filter_files.with_llm(self.index_manager.index_filter_llm).with_meta(
+                        meta_holder).with_return_type(FileNumberList).run(chunk, self.args.query)
+                    end_time = time.monotonic()
+                    
+                    total_input_cost = 0.0
+                    total_output_cost = 0.0
+                    if meta_holder.get_meta():
+                        meta_dict = meta_holder.get_meta()                    
+                        total_input_cost = meta_dict.get("input_tokens_count", 0) * model_info_map.get(model_name, {}).get("input_price", 0.0) / 1000000
+                        total_output_cost = meta_dict.get("generated_tokens_count", 0) * model_info_map.get(model_name, {}).get("output_price", 0.0) / 1000000
+                    
+                    self.printer.print_in_terminal(
+                        "quick_filter_stats",
+                        style="blue",
+                        input_tokens=meta_dict.get("input_tokens_count", 0),
+                        output_tokens=meta_dict.get("generated_tokens_count", 0),
+                        input_cost=total_input_cost,
+                        output_cost=total_output_cost,
+                        model_names=model_name,
+                        elapsed_time=f"{end_time - start_time:.2f}"
+                    )
+
                 if file_number_list:
                     for file_number in file_number_list.file_list:
-                        file_path = get_file_path(chunk[file_number].module_name)
+                        file_path = get_file_path(
+                            chunk[file_number].module_name)
                         files[file_path] = TargetFile(
                             file_path=chunk[file_number].module_name,
-                            reason=self.printer.get_message_from_key("quick_filter_reason")
+                            reason=self.printer.get_message_from_key(
+                                "quick_filter_reason")
                         )
                 return QuickFilterResult(
                     files=files,
                     has_error=False
                 )
-                            
+
             except Exception as e:
                 self.printer.print_in_terminal(
                     "quick_filter_failed",
@@ -123,25 +194,25 @@ class QuickFilter():
         if chunks:
             with ThreadPoolExecutor() as executor:
                 # 提交所有chunks到线程池并收集结果
-                futures = [executor.submit(process_chunk, i, chunk) 
-                          for i, chunk in enumerate(chunks)]
-                
+                futures = [executor.submit(process_chunk, i, chunk)
+                           for i, chunk in enumerate(chunks)]
+
                 # 等待所有任务完成并收集结果
                 for future in futures:
                     results.append(future.result())
-        
+
         # 合并所有结果
         final_files: Dict[str, TargetFile] = {}
         has_error = False
         error_messages: List[str] = []
-        
+
         for result in results:
             if result.has_error:
                 has_error = True
                 if result.error_message:
                     error_messages.append(result.error_message)
             final_files.update(result.files)
-        
+
         return QuickFilterResult(
             files=final_files,
             has_error=has_error,
@@ -149,7 +220,7 @@ class QuickFilter():
         )
 
     @byzerllm.prompt()
-    def quick_filter_files(self,file_meta_list:List[IndexItem],query:str) -> str:
+    def quick_filter_files(self, file_meta_list: List[IndexItem], query: str) -> str:
         '''
         当用户提一个需求的时候，我们需要找到相关的文件，然后阅读这些文件，并且修改其中部分文件。
         现在，给定下面的索引文件：
@@ -160,7 +231,7 @@ class QuickFilter():
 
         索引文件包含文件序号(##[]括起来的部分)，文件路径，文件符号信息等。
         下面是用户的查询需求：
-        
+
         <query>
         {{ query }}
         </query>
@@ -182,63 +253,101 @@ class QuickFilter():
         2. 如果 query 里是一段历史对话，那么对话里的内容提及的文件路径必须要返回。
         3. json格式数据不允许有注释        
         '''
-        file_meta_str = "\n".join([f"##[{index}]{item.module_name}\n{item.symbols}" for index,item in enumerate(file_meta_list)])
+        file_meta_str = "\n".join(
+            [f"##[{index}]{item.module_name}\n{item.symbols}" for index, item in enumerate(file_meta_list)])
         context = {
             "content": file_meta_str,
             "query": query
         }
-        return context    
+        return context
 
     def filter(self, index_items: List[IndexItem], query: str) -> QuickFilterResult:
         final_files: Dict[str, TargetFile] = {}
-        start_time = time.monotonic()        
+        start_time = time.monotonic()
+
+        prompt_str = self.quick_filter_files.prompt(index_items, query)
+
+        tokens_len = count_tokens(prompt_str)
 
-        prompt_str = self.quick_filter_files.prompt(index_items,query)            
-        
-        tokens_len = count_tokens(prompt_str)            
-        
         # Print current index size
         self.printer.print_in_terminal(
             "quick_filter_tokens_len",
             style="blue",
             tokens_len=tokens_len
         )
-        
-        if tokens_len > self.max_tokens:                        
+
+        if tokens_len > self.max_tokens:
             return self.big_filter(index_items)
-        
+
         try:
-            model_name = ",".join(get_llm_names(self.index_manager.index_filter_llm))
-            
+            # 获取模型名称
+            model_names = get_llm_names(self.index_manager.index_filter_llm)
+            model_name = ",".join(model_names)
+
+            # 获取模型价格信息
+            model_info_map = {}
+            for name in model_names:
+                # 第二个参数是产品模式,从args中获取
+                info = get_model_info(name, self.args.product_mode)
+                if info:
+                    model_info_map[name] = {
+                        # 每百万tokens成本
+                        "input_price": info.get("input_price", 0.0),
+                        # 每百万tokens成本
+                        "output_price": info.get("output_price", 0.0)
+                    }
+
             # 渲染 Prompt 模板
-            query = self.quick_filter_files.prompt(index_items, self.args.query)
-            
+            query = self.quick_filter_files.prompt(
+                index_items, self.args.query)
+
             # 使用流式输出处理
             stream_generator = stream_chat_with_continue(
                 self.index_manager.index_filter_llm,
                 [{"role": "user", "content": query}],
                 {}
             )
-            
+
             # 获取完整响应
             full_response, last_meta = stream_out(
                 stream_generator,
                 model_name=model_name,
-                title=self.printer.get_message_from_key_with_format("quick_filter_title", model_name=model_name),
+                title=self.printer.get_message_from_key_with_format(
+                    "quick_filter_title", model_name=model_name),
                 args=self.args
-            )                
+            )
             # 解析结果
             file_number_list = to_model(full_response, FileNumberList)
-            end_time = time.monotonic()            
-            # 打印 token 统计信息
+            end_time = time.monotonic()
+
+            # 计算总成本
+            total_input_cost = 0.0
+            total_output_cost = 0.0
+
+            for name in model_names:
+                info = model_info_map.get(name, {})
+                # 计算公式:token数 * 单价 / 1000000
+                total_input_cost += (last_meta.input_tokens_count *
+                                     info.get("input_price", 0.0)) / 1000000
+                total_output_cost += (last_meta.generated_tokens_count *
+                                      info.get("output_price", 0.0)) / 1000000
+
+            # 四舍五入到4位小数
+            total_input_cost = round(total_input_cost, 4)
+            total_output_cost = round(total_output_cost, 4)
+
+            # 打印 token 统计信息和成本
             self.printer.print_in_terminal(
-                "quick_filter_stats", 
+                "quick_filter_stats",
                 style="blue",
                 elapsed_time=f"{end_time - start_time:.2f}",
                 input_tokens=last_meta.input_tokens_count,
-                output_tokens=last_meta.generated_tokens_count
+                output_tokens=last_meta.generated_tokens_count,
+                input_cost=total_input_cost,
+                output_cost=total_output_cost,
+                model_names=model_name
             )
-            
+
         except Exception as e:
             self.printer.print_in_terminal(
                 "quick_filter_failed",
@@ -250,16 +359,17 @@ class QuickFilter():
                 has_error=True,
                 error_message=str(e)
             )
-        
+
         if file_number_list:
             for file_number in file_number_list.file_list:
                 final_files[get_file_path(index_items[file_number].module_name)] = TargetFile(
                     file_path=index_items[file_number].module_name,
-                    reason=self.printer.get_message_from_key("quick_filter_reason")
+                    reason=self.printer.get_message_from_key(
+                        "quick_filter_reason")
                 )
-        end_time = time.monotonic()            
-        self.stats["timings"]["quick_filter"] = end_time - start_time            
+        end_time = time.monotonic()
+        self.stats["timings"]["quick_filter"] = end_time - start_time
         return QuickFilterResult(
             files=final_files,
             has_error=False
-        )    
+        )

autocoder/models.py

@@ -127,11 +127,23 @@ def update_model_input_price(name: str, price: float) -> bool:
     """更新模型输入价格
     
     Args:
-        name: 模型名称
-        price: 输入价格(M/百万input tokens)
+        name (str): 要更新的模型名称，必须与models.json中的记录匹配
+        price (float): 新的输入价格，单位：美元/百万tokens。必须大于等于0
         
     Returns:
-        bool: 是否更新成功
+        bool: 是否成功找到并更新了模型价格
+        
+    Raises:
+        ValueError: 如果price为负数时抛出
+        
+    Example:
+        >>> update_model_input_price("gpt-4", 3.0)
+        True
+        
+    Notes:
+        1. 价格设置后会立即生效并保存到models.json
+        2. 实际费用计算时会按实际使用量精确到小数点后6位
+        3. 设置价格为0表示该模型当前不可用
     """
     if price < 0:
         raise ValueError("Price cannot be negative")
@@ -151,11 +163,23 @@ def update_model_output_price(name: str, price: float) -> bool:
     """更新模型输出价格
     
     Args:
-        name: 模型名称
-        price: 输出价格(M/百万output tokens)
+        name (str): 要更新的模型名称，必须与models.json中的记录匹配
+        price (float): 新的输出价格，单位：美元/百万tokens。必须大于等于0
         
     Returns:
-        bool: 是否更新成功
+        bool: 是否成功找到并更新了模型价格
+        
+    Raises:
+        ValueError: 如果price为负数时抛出
+        
+    Example:
+        >>> update_model_output_price("gpt-4", 6.0)
+        True
+        
+    Notes:
+        1. 输出价格通常比输入价格高30%-50%
+        2. 对于按token计费的API，实际收费按(input_tokens * input_price + output_tokens * output_price)计算
+        3. 价格变更会影响所有依赖模型计费的功能（如成本预测、用量监控等）
     """
     if price < 0:
         raise ValueError("Price cannot be negative")

autocoder/pyproject/__init__.py

@@ -116,6 +116,7 @@ class PyProject:
             "actions",
             ".vscode",
             ".idea",
+            "venv",
         ]
 
     @byzerllm.prompt()

autocoder/suffixproject/__init__.py

@@ -56,6 +56,7 @@ class SuffixProject:
             ".vscode",
             "actions",
             ".idea",
+            "venv",
         ]
 
     @byzerllm.prompt()

autocoder/tsproject/__init__.py

@@ -48,6 +48,7 @@ class TSProject:
             "actions",
             ".vscode",
             ".idea",
+            "venv",
         ]
 
     @byzerllm.prompt()

autocoder/utils/llms.py

@@ -3,9 +3,15 @@ from typing import Union,Optional
 
 def get_llm_names(llm: Union[byzerllm.ByzerLLM, byzerllm.SimpleByzerLLM,str],target_model_type:Optional[str]=None):
    if target_model_type is None:
+      if isinstance(llm,list):
+         return [_llm.default_model_name for _llm in llm]
       return [llm.default_model_name for llm in [llm] if llm.default_model_name]
+   
    llms = llm.get_sub_client(target_model_type) 
+   
    if llms is None:
+      if isinstance(llm,list):
+         return [_llm.default_model_name for _llm in llm]
       return [llm.default_model_name for llm in [llm] if llm.default_model_name]
    elif isinstance(llms, list):
        return [llm.default_model_name for llm in llms if llm.default_model_name]
@@ -14,6 +20,27 @@ def get_llm_names(llm: Union[byzerllm.ByzerLLM, byzerllm.SimpleByzerLLM,str],tar
    else:
       return [llm.default_model_name for llm in [llms] if llm.default_model_name]
 
+def get_model_info(model_names: str, product_mode: str):    
+    from autocoder import models as models_module
+    def get_model_by_name(model_name: str):
+        try:
+            return models_module.get_model_by_name(model_name)
+        except Exception as e:
+            return None
+    
+    if product_mode == "pro":
+        return None
+
+    if product_mode == "lite":
+        if "," in model_names:
+            # Multiple code models specified
+            model_names = model_names.split(",")
+            for _, model_name in enumerate(model_names):
+                return get_model_by_name(model_name)
+        else:
+            # Single code model
+            return get_model_by_name(model_names)
+
 def get_single_llm(model_names: str, product_mode: str):
     from autocoder import models as models_module
     if product_mode == "pro":

autocoder/version.py

@@ -1 +1 @@
-__version__ = "0.1.255"
+__version__ = "0.1.256"