Feature/basic proxy (#4164)

* basic proxy (#54)

* basic proxy (#55)

* basic proxy

* basic proxy

* basic proxy

* basic proxy

* add back double quotes

* add more specific files

* write file

* basic proxy

* Put back double quotes

.github/workflows/add-cassettes.yml

@@ -0,0 +1,49 @@
+name: Merge and Commit Cassettes
+
+on:
+  pull_request_target:
+    types:
+      - closed
+
+jobs:
+  update-cassettes:
+    if: github.event.pull_request.merged == true
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v3
+        with:
+          fetch-depth: 0  # This is necessary to fetch all branches and tags
+
+      - name: Fetch all branches
+        run: git fetch --all
+
+      - name: Reset branch
+        run: |
+          git checkout ${{ github.event.pull_request.base.ref }}
+          git reset --hard origin/cassette-diff-${{ github.event.pull_request.number }}
+
+      - name: Create PR
+        id: create_pr
+        uses: peter-evans/create-pull-request@v5
+        with:
+          commit-message: Update cassettes
+          signoff: false
+          branch: cassette-diff-${{ github.event.pull_request.number }}
+          delete-branch: false
+          title: "Update cassettes"
+          body: "This PR updates the cassettes."
+          draft: false
+
+      - name: Check PR
+        run: |
+          echo "Pull Request Number - ${{ steps.create_pr.outputs.pull-request-number }}"
+          echo "Pull Request URL - ${{ steps.create_pr.outputs.pull-request-url }}"
+
+      - name: Comment PR URL in the current PR
+        uses: thollander/actions-comment-pull-request@v2
+        with:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          message: |
+            New pull request created for cassettes: [HERE](${{ steps.create_pr.outputs.pull-request-url }}). Please merge it asap.

.github/workflows/ci.yml

@@ -3,12 +3,12 @@ name: Python CI
 on:
   push:
     branches: [ master ]
-  pull_request:
+  pull_request_target:
     branches: [ master, stable ]
 
 concurrency:
   group: ${{ format('ci-{0}', github.head_ref && format('pr-{0}', github.event.pull_request.number) || github.sha) }}
-  cancel-in-progress: ${{ github.event_name == 'pull_request' }}
+  cancel-in-progress: ${{ github.event_name == 'pull_request_target' }}
 
 jobs:
   lint:
@@ -19,6 +19,9 @@ jobs:
     steps:
       - name: Checkout repository
         uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+          ref: ${{ github.event.pull_request.head.ref }}
 
       - name: Set up Python ${{ env.min-python-version }}
         uses: actions/setup-python@v2
@@ -58,6 +61,9 @@ jobs:
     steps:
       - name: Check out repository
         uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+          ref: ${{ github.event.pull_request.head.ref }}
 
       - name: Set up Python ${{ matrix.python-version }}
         uses: actions/setup-python@v2
@@ -74,6 +80,20 @@ jobs:
           pytest -n auto --cov=autogpt --cov-report term-missing --cov-branch --cov-report xml --cov-report term
         env:
           CI: true
+          PROXY: ${{ vars.PROXY }}
+          AGENT_MODE: ${{ vars.AGENT_MODE }}
+          AGENT_TYPE: ${{ vars.AGENT_TYPE }}
 
       - name: Upload coverage reports to Codecov
         uses: codecov/codecov-action@v3
+
+      - name: Stage new files and commit
+        run: |
+          git add tests
+          git diff --cached --quiet && echo "No changes to commit" && exit 0
+          git config user.email "github-actions@github.com"
+          git config user.name "GitHub Actions"
+          git commit -m "Add new cassettes"
+          git checkout -b cassette-diff-${{ github.event.pull_request.number }}
+          git push -f origin cassette-diff-${{ github.event.pull_request.number }}
+          echo "COMMIT_SHA=$(git rev-parse HEAD)" >> $GITHUB_ENV

autogpt/llm/api_manager.py

@@ -59,10 +59,11 @@ def create_chat_completion(
                 max_tokens=max_tokens,
                 api_key=cfg.openai_api_key,
             )
-        logger.debug(f"Response: {response}")
-        prompt_tokens = response.usage.prompt_tokens
-        completion_tokens = response.usage.completion_tokens
-        self.update_cost(prompt_tokens, completion_tokens, model)
+        if not hasattr(response, "error"):
+            logger.debug(f"Response: {response}")
+            prompt_tokens = response.usage.prompt_tokens
+            completion_tokens = response.usage.completion_tokens
+            self.update_cost(prompt_tokens, completion_tokens, model)
         return response
 
     def update_cost(self, prompt_tokens, completion_tokens, model):

autogpt/llm/llm_utils.py

@@ -181,7 +181,7 @@ def create_chat_completion(
                 )
                 warned_user = True
         except (APIError, Timeout) as e:
-            if e.http_status != 502:
+            if e.http_status != 502 :
                 raise
             if attempt == num_retries - 1:
                 raise

tests/conftest.py

@@ -1,3 +1,4 @@
+import os
 from pathlib import Path
 
 import pytest
@@ -9,6 +10,8 @@
 
 pytest_plugins = ["tests.integration.agent_factory"]
 
+PROXY = os.environ.get("PROXY")
+
 
 @pytest.fixture()
 def workspace_root(tmp_path: Path) -> Path:

tests/integration/challenges/information_retrieval/test_information_retrieval_challenge_a.py

@@ -26,7 +26,7 @@ def input_generator(input_sequence: list) -> Generator[str, None, None]:
 @requires_api_key("OPENAI_API_KEY")
 @run_multiple_times(3)
 def test_information_retrieval_challenge_a(
-    get_company_revenue_agent, monkeypatch
+    get_company_revenue_agent, monkeypatch, patched_api_requestor
 ) -> None:
     """
     Test the challenge_a function in a given agent by mocking user inputs and checking the output file content.

tests/integration/challenges/memory/test_memory_challenge_a.py

@@ -13,7 +13,7 @@
 @pytest.mark.vcr
 @requires_api_key("OPENAI_API_KEY")
 def test_memory_challenge_a(
-    memory_management_agent: Agent, user_selected_level: int
+    memory_management_agent: Agent, user_selected_level: int, patched_api_requestor
 ) -> None:
     """
     The agent reads a file containing a task_id. Then, it reads a series of other files.
@@ -30,7 +30,7 @@ def test_memory_challenge_a(
     create_instructions_files(memory_management_agent, num_files, task_id)
 
     try:
-        run_interaction_loop(memory_management_agent, 180)
+        run_interaction_loop(memory_management_agent, 400)
     # catch system exit exceptions
     except SystemExit:
         file_path = str(memory_management_agent.workspace.get_path("output.txt"))

tests/integration/challenges/memory/test_memory_challenge_b.py

@@ -14,7 +14,7 @@
 @pytest.mark.vcr
 @requires_api_key("OPENAI_API_KEY")
 def test_memory_challenge_b(
-    memory_management_agent: Agent, user_selected_level: int
+    memory_management_agent: Agent, user_selected_level: int, patched_api_requestor
 ) -> None:
     """
     The agent reads a series of files, each containing a task_id and noise. After reading 'n' files,

tests/integration/challenges/memory/test_memory_challenge_c.py

@@ -14,7 +14,7 @@
 @pytest.mark.vcr
 @requires_api_key("OPENAI_API_KEY")
 def test_memory_challenge_c(
-    memory_management_agent: Agent, user_selected_level: int
+    memory_management_agent: Agent, user_selected_level: int, patched_api_requestor
 ) -> None:
     """
     Instead of reading task Ids from files as with the previous challenges, the agent now must remember

tests/integration/conftest.py

@@ -1,7 +1,9 @@
 import os
 
+import openai
 import pytest
 
+from tests.conftest import PROXY
 from tests.vcr.vcr_filter import before_record_request, before_record_response
 
 
@@ -17,5 +19,37 @@ def vcr_config():
             "X-OpenAI-Client-User-Agent",
             "User-Agent",
         ],
-        "match_on": ["method", "uri", "body"],
+        "match_on": ["method", "body"],
     }
+
+
+def patch_api_base(requestor):
+    new_api_base = f"{PROXY}/v1"
+    requestor.api_base = new_api_base
+    return requestor
+
+
+@pytest.fixture
+def patched_api_requestor(mocker):
+    original_init = openai.api_requestor.APIRequestor.__init__
+    original_validate_headers = openai.api_requestor.APIRequestor._validate_headers
+
+    def patched_init(requestor, *args, **kwargs):
+        original_init(requestor, *args, **kwargs)
+        patch_api_base(requestor)
+
+    def patched_validate_headers(self, supplied_headers):
+        headers = original_validate_headers(self, supplied_headers)
+        headers["AGENT-MODE"] = os.environ.get("AGENT_MODE")
+        headers["AGENT-TYPE"] = os.environ.get("AGENT_TYPE")
+        return headers
+
+    if PROXY:
+        mocker.patch("openai.api_requestor.APIRequestor.__init__", new=patched_init)
+        mocker.patch.object(
+            openai.api_requestor.APIRequestor,
+            "_validate_headers",
+            new=patched_validate_headers,
+        )
+
+    return mocker

tests/integration/goal_oriented/cassettes/test_write_file/test_write_file.yaml

@@ -573,4 +573,120 @@ interactions:
     status:
       code: 200
       message: OK
+- request:
+    body: '{"model": "gpt-3.5-turbo", "messages": [{"role": "system", "content": "You
+      are write_to_file-GPT, an AI designed to use the write_to_file command to write
+      ''Hello World'' into a file named \"hello_world.txt\" and then use the task_complete
+      command to complete the task.\nYour decisions must always be made independently
+      without seeking user assistance. Play to your strengths as an LLM and pursue
+      simple strategies with no legal complications.\n\nGOALS:\n\n1. Use the write_to_file
+      command to write ''Hello World'' into a file named \"hello_world.txt\".\n2.
+      Use the task_complete command to complete the task.\n3. Do not use any other
+      commands.\n\n\nConstraints:\n1. ~4000 word limit for short term memory. Your
+      short term memory is short, so immediately save important information to files.\n2.
+      If you are unsure how you previously did something or want to recall past events,
+      thinking about similar events will help you remember.\n3. No user assistance\n4.
+      Exclusively use the commands listed in double quote e.g. \"command name\"\n\nCommands:\n1.
+      append_to_file: Append to file, args: \"filename\": \"<filename>\", \"text\":
+      \"<text>\"\n2. delete_file: Delete file, args: \"filename\": \"<filename>\"\n3.
+      list_files: List Files in Directory, args: \"directory\": \"<directory>\"\n4.
+      read_file: Read file, args: \"filename\": \"<filename>\"\n5. write_to_file:
+      Write to file, args: \"filename\": \"<filename>\", \"text\": \"<text>\"\n6.
+      delete_agent: Delete GPT Agent, args: \"key\": \"<key>\"\n7. get_hyperlinks:
+      Get text summary, args: \"url\": \"<url>\"\n8. get_text_summary: Get text summary,
+      args: \"url\": \"<url>\", \"question\": \"<question>\"\n9. list_agents: List
+      GPT Agents, args: () -> str\n10. message_agent: Message GPT Agent, args: \"key\":
+      \"<key>\", \"message\": \"<message>\"\n11. start_agent: Start GPT Agent, args:
+      \"name\": \"<name>\", \"task\": \"<short_task_desc>\", \"prompt\": \"<prompt>\"\n12.
+      task_complete: Task Complete (Shutdown), args: \"reason\": \"<reason>\"\n\nResources:\n1.
+      Internet access for searches and information gathering.\n2. Long Term memory
+      management.\n3. GPT-3.5 powered Agents for delegation of simple tasks.\n4. File
+      output.\n\nPerformance Evaluation:\n1. Continuously review and analyze your
+      actions to ensure you are performing to the best of your abilities.\n2. Constructively
+      self-criticize your big-picture behavior constantly.\n3. Reflect on past decisions
+      and strategies to refine your approach.\n4. Every command has a cost, so be
+      smart and efficient. Aim to complete tasks in the least number of steps.\n5.
+      Write all code to a file.\n\nYou should only respond in JSON format as described
+      below \nResponse Format: \n{\n    \"thoughts\": {\n        \"text\": \"thought\",\n        \"reasoning\":
+      \"reasoning\",\n        \"plan\": \"- short bulleted\\n- list that conveys\\n-
+      long-term plan\",\n        \"criticism\": \"constructive self-criticism\",\n        \"speak\":
+      \"thoughts summary to say to user\"\n    },\n    \"command\": {\n        \"name\":
+      \"command name\",\n        \"args\": {\n            \"arg name\": \"value\"\n        }\n    }\n}
+      \nEnsure the response can be parsed by Python json.loads"}, {"role": "system",
+      "content": "The current time and date is Tue Jan  1 00:00:00 2000"}, {"role":
+      "user", "content": "Determine which next command to use, and respond using the
+      format specified above:"}], "temperature": 0, "max_tokens": 0}'
+    headers:
+      Accept:
+      - '*/*'
+      Accept-Encoding:
+      - gzip, deflate
+      Connection:
+      - keep-alive
+      Content-Length:
+      - '3405'
+      Content-Type:
+      - application/json
+    method: POST
+    uri: https://api.openai.com/v1/chat/completions
+  response:
+    body:
+      string: !!binary |
+        H4sIAAAAAAAAA7yTT4/TMBDF73yK0Vx6cauUqt1trmhhK8QBBEKIoMrrTBvT2BPsCVu2yndfJemf
+        3SBOCK7zxu/9xh4f0OaYoim0GFeV46vX7u3i183N/Geup/7Lq9vVh/fXEt49PLwpNCrku+9k5Hhi
+        YthVJYlljwpNIC2UYzpdXM+Wy3kySxQ6zqnEFLeVjGeT+VjqcMfjZJZMUWEd9ZYwPWAV2FWyFt6R
+        j5heLRKFF+9z/WUyVSgsujyXlotpo9AUbA1FTL8e0FE82QYuCVPUMdoo2ksLyV7ItwMcMg8AkKEU
+        XG8LiRmmcCweBdpLW8xwBZ4oB2GoI4EUBPfBCq2F1xtbEhh2TvuuoRNgdEtlyfCZQ5mPwHph0NC1
+        eu0oh1HR6uv7Vp/IXkaTDNXT7EA6srd+2wN8LAhExx0E+lHbQBEc/UWago72j3PY2ImOo4CuqsBV
+        sFoINhxAilbVcTdErkrte9oxfPpP12SCFWtsdMN3Ih/r0DJogdX51QyHQEYuEf090F4uTMJda9sy
+        TIsV6d0p6d6W5b9chz64Uac1PZr+tqWtQ8/0DGKArsN2uOC90PZeLAYcz0yGn+LJTCfajvgInvkG
+        G4Ub620s1v0+Y4pRuEKF1ue0xzRpvjUvHgEAAP//AwDSj7qBhAQAAA==
+    headers:
+      CF-Cache-Status:
+      - DYNAMIC
+      CF-RAY:
+      - 7c6c3f8bcdd1cf87-SJC
+      Cache-Control:
+      - no-cache, must-revalidate
+      Connection:
+      - keep-alive
+      Content-Encoding:
+      - gzip
+      Content-Type:
+      - application/json
+      Date:
+      - Sat, 13 May 2023 16:24:06 GMT
+      Server:
+      - cloudflare
+      access-control-allow-origin:
+      - '*'
+      alt-svc:
+      - h3=":443"; ma=86400, h3-29=":443"; ma=86400
+      openai-model:
+      - gpt-3.5-turbo-0301
+      openai-organization:
+      - user-adtx4fhfg1qsiyzdoaxciooj
+      openai-processing-ms:
+      - '16269'
+      openai-version:
+      - '2020-10-01'
+      strict-transport-security:
+      - max-age=15724800; includeSubDomains
+      x-ratelimit-limit-requests:
+      - '3500'
+      x-ratelimit-limit-tokens:
+      - '90000'
+      x-ratelimit-remaining-requests:
+      - '3499'
+      x-ratelimit-remaining-tokens:
+      - '86496'
+      x-ratelimit-reset-requests:
+      - 17ms
+      x-ratelimit-reset-tokens:
+      - 2.336s
+      x-request-id:
+      - 8d3e6826e88e77fb2cbce01166ddc550
+    status:
+      code: 200
+      message: OK
 version: 1

tests/integration/goal_oriented/test_browse_website.py

@@ -8,7 +8,7 @@
 
 @requires_api_key("OPENAI_API_KEY")
 @pytest.mark.vcr
-def test_browse_website(browser_agent: Agent) -> None:
+def test_browse_website(browser_agent: Agent, patched_api_requestor) -> None:
     file_path = browser_agent.workspace.get_path("browse_website.txt")
     try:
         run_interaction_loop(browser_agent, 120)

tests/integration/goal_oriented/test_write_file.py

@@ -1,3 +1,6 @@
+import os
+
+import openai
 import pytest
 
 from autogpt.agent import Agent
@@ -8,10 +11,10 @@
 
 @requires_api_key("OPENAI_API_KEY")
 @pytest.mark.vcr
-def test_write_file(writer_agent: Agent) -> None:
+def test_write_file(writer_agent: Agent, patched_api_requestor) -> None:
     file_path = str(writer_agent.workspace.get_path("hello_world.txt"))
     try:
-        run_interaction_loop(writer_agent, 40)
+        run_interaction_loop(writer_agent, 200)
     # catch system exit exceptions
     except SystemExit:  # the agent returns an exception when it shuts down
         content = read_file(file_path)