Other small fixes (#61)

ml6team · Jan 9, 2024 · f5a4b1c · f5a4b1c
1 parent dc221f3
commit f5a4b1c
Show file tree

Hide file tree

Showing 5 changed files with 22 additions and 12 deletions.
diff --git a/src/evaluation.ipynb b/src/evaluation.ipynb
@@ -407,10 +407,12 @@
     "import os\n",
     "import pipeline_eval\n",
     "\n",
+    "os.environ[\"OPENAI_KEY\"] = \"\"\n",
+    "\n",
     "evaluation_args = {\n",
     "    \"retrieval_top_k\": 2,\n",
-    "    \"evaluation_module\": \"langchain.llms\",\n",
-    "    \"evaluation_llm\": \"OpenAI\",\n",
+    "    \"evaluation_module\": \"langchain.chat_models\",\n",
+    "    \"evaluation_llm\": \"ChatOpenAI\",\n",
     "    \"evaluation_llm_kwargs\": {\"openai_api_key\": os.environ[\"OPENAI_KEY\"], \"model_name\" : \"gpt-3.5-turbo\"},  # TODO: Update with your key or use a different model\n",
     "    \"evaluation_metrics\": [\"context_precision\", \"context_relevancy\"]\n",
     "}\n",
@@ -433,8 +435,15 @@
    "source": [
     "runner = DockerRunner()\n",
     "extra_volumes = [str(os.path.join(os.path.abspath('.'), \"evaluation_datasets\")) + \":/data\"]\n",
-    "runner.run(evaluation_pipeline, extra_volumes=extra_volumes)\n",
-    "\n",
+    "runner.run(evaluation_pipeline, extra_volumes=extra_volumes)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
     "utils.store_results(evaluation_pipeline.name, **shared_args, **indexing_args)"
    ]
   },

diff --git a/src/parameter_search.ipynb b/src/parameter_search.ipynb
@@ -216,7 +216,7 @@
     "import weaviate\n",
     "\n",
     "try:\n",
-    "    local_weaviate_client = weaviate.Client(\"http://localhost:8080\")\n",
+    "    local_weaviate_client = weaviate.Client(\"http://localhost:8081\")\n",
     "    logging.info(\"Connected to Weaviate instance\")\n",
     "except weaviate.WeaviateStartUpError:\n",
     "    logging.error(\"Cannot connect to weaviate instance, is it running?\")"
@@ -324,17 +324,17 @@
     "\n",
     "fixed_args = {\n",
     "    \"base_path\":BASE_PATH,\n",
-    "    \"weaviate_url\":f\"http://{host_ip}:8080\", # IP address \n",
+    "    \"weaviate_url\":f\"http://{host_ip}:8081\", # IP address \n",
     "}\n",
     "fixed_index_args = {\n",
     "    \"n_rows_to_load\":1000,\n",
     "}\n",
     "fixed_eval_args = {\n",
     "    \"csv_dataset_uri\":\"/data/wikitext_1000_q.csv\", #make sure it is the same as mounted file\n",
     "    \"csv_separator\":\";\",\n",
-    "    \"evaluation_module\": \"langchain.llms\",\n",
-    "    \"evaluation_llm\":\"OpenAI\",\n",
-    "    \"evaluation_llm_kwargs\":{\"openai_api_key\": \"\"}, #TODO Specify your key if you're using OpenAI\n",
+    "    \"evaluation_module\": \"langchain.chat_models\",\n",
+    "    \"evaluation_llm\": \"ChatOpenAI\",\n",
+    "    \"evaluation_llm_kwargs\": {\"openai_api_key\": os.environ[\"OPENAI_KEY\"], \"model_name\" : \"gpt-3.5-turbo\"},  # TODO: Update with your key or use a different model\n",
     "    \"evaluation_metrics\":[\"context_precision\", \"context_relevancy\"]\n",
     "}"
    ]

diff --git a/src/pipeline.ipynb b/src/pipeline.ipynb
@@ -288,7 +288,8 @@
    "source": [
     "## Exploring the dataset\n",
     "\n",
-    "You can also explore the dataset using the fondant explorer, this enables you to visualize your output dataset at each component step. It might take a while to start the first time as it needs to download the explorer docker image first."
+    "You can also explore the dataset using the fondant explorer, this enables you to visualize your output dataset at each component step. It might take a while to start the first time as it needs to download the explorer docker image first. You can browse at \n",
+    "http://localhost:8501/"
    ]
   },
   {

diff --git a/src/pipeline_eval.py b/src/pipeline_eval.py
@@ -6,8 +6,8 @@
 
 def create_pipeline(
     *,
+    weaviate_url: str,
     base_path: str = "./data-dir",
-    weaviate_url="http://host.docker.internal:8080",
     weaviate_class: str = "Pipeline1",
     csv_dataset_uri: str = "/data/wikitext_1000_q.csv",
     csv_separator: str = ";",

diff --git a/src/pipeline_index.py b/src/pipeline_index.py
@@ -5,9 +5,9 @@
 
 def create_pipeline(
     *,
+    weaviate_url: str,
     base_path: str = "./data-dir",
     n_rows_to_load: int = 1000,
-    weaviate_url: str = "http://host.docker.internal:8080",
     weaviate_class: str = "Pipeline1",
     weaviate_overwrite: bool = True,
     embed_model_provider: str = "huggingface",