Hallacy/11 4 release (#54)

* Make embeddings_utils be importable (#104)

* Make embeddings_utils be importable

* Small tweaks to dicts for typing

* Remove default api_prefix and move v1 prefix to default api_base (#95)

* make construct_from key argument optional (#92)

* Split search.prepare_data into answers/classifications/search versions (#93)

* Break out prepare_data into answers, classifications, and search

* And cleaned up CLI

* Validate search files (#69)

* Add validators for search files

* Clean up fields

Co-authored-by: kennyhsu5 <[email protected]>
Co-authored-by: Madeleine Thompson <[email protected]>

Author: 3 people

examples/embeddings/Classification.ipynb

@@ -90,7 +90,7 @@
     }
    ],
    "source": [
-    "from utils import plot_multiclass_precision_recall\n",
+    "from openai.embeddings_utils import plot_multiclass_precision_recall\n",
     "\n",
     "plot_multiclass_precision_recall(probas, y_test, [1,2,3,4,5], clf)"
    ]

examples/embeddings/Code_search.ipynb

@@ -185,7 +185,7 @@
     }
    ],
    "source": [
-    "from utils import get_embedding\n",
+    "from openai.embeddings_utils import get_embedding\n",
     "\n",
     "df = pd.DataFrame(all_funcs)\n",
     "df['code_embedding'] = df['code'].apply(lambda x: get_embedding(x, engine='babbage-code-search-code'))\n",
@@ -231,7 +231,7 @@
     }
    ],
    "source": [
-    "from utils import cosine_similarity\n",
+    "from openai.embeddings_utils import cosine_similarity\n",
     "\n",
     "def search_functions(df, code_query, n=3, pprint=True, n_lines=7):\n",
     "    embedding = get_embedding(code_query, engine='babbage-code-search-text')\n",

examples/embeddings/Obtain_dataset.ipynb

@@ -156,7 +156,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "from utils import get_embedding\n",
+    "from openai.embeddings_utils import get_embedding\n",
     "\n",
     "# This will take just under 10 minutes\n",
     "df['babbage_similarity'] = df.combined.apply(lambda x: get_embedding(x, engine='babbage-similarity'))\n",

examples/embeddings/Semantic_text_search_using_embeddings.ipynb

@@ -49,7 +49,7 @@
     }
    ],
    "source": [
-    "from utils import get_embedding, cosine_similarity\n",
+    "from openai.embeddings_utils import get_embedding, cosine_similarity\n",
     "\n",
     "# search through the reviews for a specific product\n",
     "def search_reviews(df, product_description, n=3, pprint=True):\n",

examples/embeddings/User_and_product_embeddings.ipynb

@@ -70,7 +70,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "from utils import cosine_similarity\n",
+    "from openai.embeddings_utils import cosine_similarity\n",
     "\n",
     "# evaluate embeddings as recommendations on X_test\n",
     "def evaluate_single_match(row):\n",

examples/embeddings/Zero-shot_classification.ipynb

@@ -78,7 +78,7 @@
     }
    ],
    "source": [
-    "from utils import cosine_similarity, get_embedding\n",
+    "from openai.embeddings_utils import cosine_similarity, get_embedding\n",
     "from sklearn.metrics import PrecisionRecallDisplay\n",
     "\n",
     "def evaluate_emeddings_approach(\n",

openai/__init__.py

@@ -25,7 +25,7 @@
 api_key_path: Optional[str] = os.environ.get("OPENAI_API_KEY_PATH")
 
 organization = os.environ.get("OPENAI_ORGANIZATION")
-api_base = os.environ.get("OPENAI_API_BASE", "https://api.openai.com")
+api_base = os.environ.get("OPENAI_API_BASE", "https://api.openai.com/v1")
 api_version = None
 verify_ssl_certs = True  # No effect. Certificates are always verified.
 proxy = None

openai/api_resources/abstract/api_resource.py

@@ -5,7 +5,7 @@
 
 
 class APIResource(OpenAIObject):
-    api_prefix = "v1"
+    api_prefix = ""
 
     @classmethod
     def retrieve(cls, id, api_key=None, request_id=None, **params):
@@ -28,7 +28,9 @@ def class_url(cls):
         # Namespaces are separated in object names with periods (.) and in URLs
         # with forward slashes (/), so replace the former with the latter.
         base = cls.OBJECT_NAME.replace(".", "/")  # type: ignore
-        return "/%s/%ss" % (cls.api_prefix, base)
+        if cls.api_prefix:
+            return "/%s/%ss" % (cls.api_prefix, base)
+        return "/%ss" % (base)
 
     def instance_url(self):
         id = self.get("id")

openai/api_resources/abstract/engine_api_resource.py

@@ -22,10 +22,10 @@ def class_url(cls, engine: Optional[str] = None):
         # with forward slashes (/), so replace the former with the latter.
         base = cls.OBJECT_NAME.replace(".", "/")  # type: ignore
         if engine is None:
-            return "/%s/%ss" % (cls.api_prefix, base)
+            return "/%ss" % (base)
 
         extn = quote_plus(engine)
-        return "/%s/engines/%s/%ss" % (cls.api_prefix, extn, base)
+        return "/engines/%s/%ss" % (extn, base)
 
     @classmethod
     def create(

openai/api_resources/answer.py

@@ -2,13 +2,11 @@
 
 
 class Answer(OpenAIObject):
-    api_prefix = "v1"
-
     @classmethod
-    def get_url(self, base):
-        return "/%s/%s" % (self.api_prefix, base)
+    def get_url(self):
+        return "/answers"
 
     @classmethod
     def create(cls, **params):
         instance = cls()
-        return instance.request("post", cls.get_url("answers"), params)
+        return instance.request("post", cls.get_url(), params)

openai/api_resources/classification.py

@@ -2,13 +2,11 @@
 
 
 class Classification(OpenAIObject):
-    api_prefix = "v1"
-
     @classmethod
-    def get_url(self, base):
-        return "/%s/%s" % (self.api_prefix, base)
+    def get_url(self):
+        return "/classifications"
 
     @classmethod
     def create(cls, **params):
         instance = cls()
-        return instance.request("post", cls.get_url("classifications"), params)
+        return instance.request("post", cls.get_url(), params)

openai/api_resources/search.py

@@ -2,14 +2,11 @@
 
 
 class Search(APIResource):
-    api_prefix = "v1"
-    OBJECT_NAME = "search_indices"
-
     @classmethod
     def class_url(cls):
-        return "/%s/%s" % (cls.api_prefix, cls.OBJECT_NAME)
+        return "/search_indices/search"
 
     @classmethod
     def create_alpha(cls, **params):
         instance = cls()
-        return instance.request("post", f"{cls.class_url()}/search", params)
+        return instance.request("post", cls.class_url(), params)

openai/cli.py

@@ -3,6 +3,7 @@
 import signal
 import sys
 import warnings
+from functools import partial
 from typing import Optional
 
 import requests
@@ -11,10 +12,12 @@
 from openai.upload_progress import BufferReader
 from openai.validators import (
     apply_necessary_remediation,
-    apply_optional_remediation,
+    apply_validators,
+    get_search_validators,
     get_validators,
     read_any_format,
     write_out_file,
+    write_out_search_file,
 )
 
 
@@ -227,6 +230,40 @@ def list(cls, args):
 
 
 class Search:
+    @classmethod
+    def prepare_data(cls, args, purpose):
+
+        sys.stdout.write("Analyzing...\n")
+        fname = args.file
+        auto_accept = args.quiet
+
+        optional_fields = ["metadata"]
+
+        if purpose == "classifications":
+            required_fields = ["text", "labels"]
+        else:
+            required_fields = ["text"]
+
+        df, remediation = read_any_format(
+            fname, fields=required_fields + optional_fields
+        )
+
+        if "metadata" not in df:
+            df["metadata"] = None
+
+        apply_necessary_remediation(None, remediation)
+        validators = get_search_validators(required_fields, optional_fields)
+
+        write_out_file_func = partial(
+            write_out_search_file,
+            purpose=purpose,
+            fields=required_fields + optional_fields,
+        )
+
+        apply_validators(
+            df, fname, remediation, validators, auto_accept, write_out_file_func
+        )
+
     @classmethod
     def create_alpha(cls, args):
         resp = openai.Search.create_alpha(
@@ -489,49 +526,14 @@ def prepare_data(cls, args):
 
         validators = get_validators()
 
-        optional_remediations = []
-        if remediation is not None:
-            optional_remediations.append(remediation)
-        for validator in validators:
-            remediation = validator(df)
-            if remediation is not None:
-                optional_remediations.append(remediation)
-                df = apply_necessary_remediation(df, remediation)
-
-        any_optional_or_necessary_remediations = any(
-            [
-                remediation
-                for remediation in optional_remediations
-                if remediation.optional_msg is not None
-                or remediation.necessary_msg is not None
-            ]
+        apply_validators(
+            df,
+            fname,
+            remediation,
+            validators,
+            auto_accept,
+            write_out_file_func=write_out_file,
         )
-        any_necessary_applied = any(
-            [
-                remediation
-                for remediation in optional_remediations
-                if remediation.necessary_msg is not None
-            ]
-        )
-        any_optional_applied = False
-
-        if any_optional_or_necessary_remediations:
-            sys.stdout.write(
-                "\n\nBased on the analysis we will perform the following actions:\n"
-            )
-            for remediation in optional_remediations:
-                df, optional_applied = apply_optional_remediation(
-                    df, remediation, auto_accept
-                )
-                any_optional_applied = any_optional_applied or optional_applied
-        else:
-            sys.stdout.write("\n\nNo remediations found.\n")
-
-        any_optional_or_necessary_applied = (
-            any_optional_applied or any_necessary_applied
-        )
-
-        write_out_file(df, fname, any_optional_or_necessary_applied, auto_accept)
 
 
 def tools_register(parser):
@@ -561,6 +563,57 @@ def help(args):
     )
     sub.set_defaults(func=FineTune.prepare_data)
 
+    sub = subparsers.add_parser("search.prepare_data")
+    sub.add_argument(
+        "-f",
+        "--file",
+        required=True,
+        help="JSONL, JSON, CSV, TSV, TXT or XLSX file containing text examples to be analyzed."
+        "This should be the local file path.",
+    )
+    sub.add_argument(
+        "-q",
+        "--quiet",
+        required=False,
+        action="store_true",
+        help="Auto accepts all suggestions, without asking for user input. To be used within scripts.",
+    )
+    sub.set_defaults(func=partial(Search.prepare_data, purpose="search"))
+
+    sub = subparsers.add_parser("classifications.prepare_data")
+    sub.add_argument(
+        "-f",
+        "--file",
+        required=True,
+        help="JSONL, JSON, CSV, TSV, TXT or XLSX file containing text-label examples to be analyzed."
+        "This should be the local file path.",
+    )
+    sub.add_argument(
+        "-q",
+        "--quiet",
+        required=False,
+        action="store_true",
+        help="Auto accepts all suggestions, without asking for user input. To be used within scripts.",
+    )
+    sub.set_defaults(func=partial(Search.prepare_data, purpose="classification"))
+
+    sub = subparsers.add_parser("answers.prepare_data")
+    sub.add_argument(
+        "-f",
+        "--file",
+        required=True,
+        help="JSONL, JSON, CSV, TSV, TXT or XLSX file containing text examples to be analyzed."
+        "This should be the local file path.",
+    )
+    sub.add_argument(
+        "-q",
+        "--quiet",
+        required=False,
+        action="store_true",
+        help="Auto accepts all suggestions, without asking for user input. To be used within scripts.",
+    )
+    sub.set_defaults(func=partial(Search.prepare_data, purpose="answer"))
+
 
 def api_register(parser):
     # Engine management