palimpzest 0.7.9__py3-none-any.whl → 0.7.11__py3-none-any.whl

palimpzest/query/generators/generators.py

@@ -215,14 +215,15 @@ class BaseGenerator(Generic[ContextType, InputType], ABC):
 
             # TODO: wrap non-list outputs in a list if expected output is a list
 
-            # common error: if the output is a singleton list which contains a list, but the expected field type
+            # common error for one-to-one: if the output is a singleton list which contains a list, but the expected field type
             # is a list of strings, or a list of floats, i.e. not a list of lists; then extract the inner list
-            for field, field_type in fields.items():
-                answer = field_answers[field]
-                field_type_is_not_list_of_lists = isinstance(field_type, ListField) and not issubclass(field_type.element_type, ListField)
-                answer_is_list_of_lists = isinstance(answer, list) and len(answer) == 1 and isinstance(answer[0], list)
-                if field_type_is_not_list_of_lists and answer_is_list_of_lists:
-                    field_answers[field] = answer[0]
+            if self.cardinality == Cardinality.ONE_TO_ONE:
+                for field, field_type in fields.items():
+                    answer = field_answers[field]
+                    field_type_is_not_list_of_lists = isinstance(field_type, ListField) and not issubclass(field_type.element_type, ListField)
+                    answer_is_list_of_lists = isinstance(answer, list) and len(answer) == 1 and isinstance(answer[0], list)
+                    if field_type_is_not_list_of_lists and answer_is_list_of_lists:
+                        field_answers[field] = answer[0]
 
             # prepare the field answers to match the expected output and return
             return self._prepare_field_answers(field_answers, fields)
@@ -416,6 +417,8 @@ class BaseGenerator(Generic[ContextType, InputType], ABC):
                 prompt += message["content"] + "\n" if message["type"] == "text" else "<image>\n"
         logger.debug(f"PROMPT:\n{prompt}")
         logger.debug(Fore.GREEN + f"{completion_text}\n" + Style.RESET_ALL)
+        print(f"PROMPT:\n{prompt}")
+        print(Fore.GREEN + f"{completion_text}\n" + Style.RESET_ALL)
 
         # parse reasoning
         reasoning = None

palimpzest/query/optimizer/rules.py

@@ -323,7 +323,8 @@ class LLMConvertBondedRule(ImplementationRule):
             first_criteria = model in pure_vision_models and not is_image_conversion
             second_criteria = model in pure_text_models and is_image_conversion
             third_criteria = model.is_llama_model() and model.is_vision_model() and (num_image_fields > 1 or list_image_field)
-            if first_criteria or second_criteria or third_criteria:
+            fourth_criteria = model.is_embedding_model()
+            if first_criteria or second_criteria or third_criteria or fourth_criteria:
                 continue
 
             # construct multi-expression
@@ -470,7 +471,7 @@ class RAGConvertRule(ImplementationRule):
         physical_expressions = []
         for model in physical_op_params["available_models"]:
             # skip this model if this is a pure image model
-            if model in pure_vision_models:
+            if model in pure_vision_models or model.is_embedding_model():
                 continue
 
             for num_chunks_per_field in cls.num_chunks_per_fields:
@@ -666,7 +667,8 @@ class CriticAndRefineConvertRule(ImplementationRule):
             first_criteria = model in pure_vision_models and not is_image_conversion
             second_criteria = model in pure_text_models and is_image_conversion
             third_criteria = model.is_llama_model() and model.is_vision_model() and (num_image_fields > 1 or list_image_field)
-            if first_criteria or second_criteria or third_criteria:
+            fourth_criteria = model.is_embedding_model()
+            if first_criteria or second_criteria or third_criteria or fourth_criteria:
                 continue
 
             models.append(model)
@@ -746,7 +748,7 @@ class SplitConvertRule(ImplementationRule):
         physical_expressions = []
         for model in physical_op_params["available_models"]:
             # skip this model if this is a pure image model
-            if model in pure_vision_models:
+            if model in pure_vision_models or model.is_embedding_model():
                 continue
 
             for min_size_to_chunk in cls.min_size_to_chunk:
@@ -934,7 +936,8 @@ class LLMFilterRule(ImplementationRule):
             first_criteria = model in pure_vision_models and not is_image_filter
             second_criteria = model in pure_text_models and is_image_filter
             third_criteria = model.is_llama_model() and model.is_vision_model() and (num_image_fields > 1 or list_image_field)
-            if first_criteria or second_criteria or third_criteria:
+            fourth_criteria = model.is_embedding_model()
+            if first_criteria or second_criteria or third_criteria or fourth_criteria:
                 continue
 
             # construct multi-expression

palimpzest/utils/model_helpers.py

@@ -25,13 +25,17 @@ def get_vision_models() -> list[Model]:
     return models
 
 
-def get_models(include_vision: bool = False) -> list[Model]:
+def get_models(include_vision: bool = False, include_embedding: bool = False) -> list[Model]:
     """
     Return the set of models which the system has access to based on the set environment variables.
     """
     models = []
     if os.getenv("OPENAI_API_KEY") is not None:
         openai_models = [model for model in Model if model.is_openai_model()]
+        if not include_embedding:
+            openai_models = [
+                model for model in openai_models if not model.is_embedding_model()
+            ]
         models.extend(openai_models)
 
     if os.getenv("TOGETHER_API_KEY") is not None:
@@ -40,6 +44,10 @@ def get_models(include_vision: bool = False) -> list[Model]:
             together_models = [
                 model for model in together_models if not model.is_vision_model()
             ]
+        if not include_embedding:
+            together_models = [
+                model for model in together_models if not model.is_embedding_model()
+            ]
         models.extend(together_models)
 
     if include_vision:

{palimpzest-0.7.9.dist-info → palimpzest-0.7.11.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: palimpzest
-Version: 0.7.9
+Version: 0.7.11
 Summary: Palimpzest is a system which enables anyone to process AI-powered analytical queries simply by defining them in a declarative language
 Author-email: MIT DSG Semantic Management Lab <michjc@csail.mit.edu>
 Project-URL: homepage, https://palimpzest.org

{palimpzest-0.7.9.dist-info → palimpzest-0.7.11.dist-info}/RECORD

@@ -36,7 +36,7 @@ palimpzest/query/execution/random_sampling_execution_strategy.py,sha256=jWPd-sBs
 palimpzest/query/execution/single_threaded_execution_strategy.py,sha256=-31gg7ha0D97g2zipzF72f7nrJFcDuiRKmo10_q5ykE,12389
 palimpzest/query/generators/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 palimpzest/query/generators/api_client_factory.py,sha256=lH9p5wRehECc7m1dWOGAGivD2F5x6hGh5kQCMw2-tfs,1065
-palimpzest/query/generators/generators.py,sha256=HOquErARyNTFya0GiudWyV2UCJEz5g1q2FufXPT7nHE,26032
+palimpzest/query/generators/generators.py,sha256=dI7KuYGeiZBAkZKhbI9hmrMPC1n2j9nZQT2__5yPUdU,26235
 palimpzest/query/operators/__init__.py,sha256=fq0YvXuqxLtQyI_LuyOe02KYqYCmD-Sh1M-FcVWnr_A,3163
 palimpzest/query/operators/aggregate.py,sha256=G7jH2k0673t0TtzchCB0a9LLhSWwZb8X20FtuNw8xTQ,10288
 palimpzest/query/operators/code_synthesis_convert.py,sha256=Oki3ZJi6lRD3NFKylulmreFniRVYjrbuKWFe_M90HXM,20992
@@ -60,7 +60,7 @@ palimpzest/query/optimizer/optimizer_strategy.py,sha256=Aek-LsQQO9VuxI1ZLkmadxik
 palimpzest/query/optimizer/optimizer_strategy_type.py,sha256=V-MMHvJdnfZKoUX1xxxwh66q1RjN2FL35IsiT1C62c8,1084
 palimpzest/query/optimizer/plan.py,sha256=i6lOTSNMcVNf0SruMD3lZ_-Nq3_UWq5X6QFd_lRvYA4,5682
 palimpzest/query/optimizer/primitives.py,sha256=r0zjavxxVCCFABXDwdQxRJmQek6WLo1777suoy7CWEc,4038
-palimpzest/query/optimizer/rules.py,sha256=lPMtOunfG-mcAXn7crYD_ULAfkBRXeeGRr8-0NzzZjU,46853
+palimpzest/query/optimizer/rules.py,sha256=msBWBleUy1sVs1iR67jMf6fPVoRaHsL438pg3nK_hHs,47141
 palimpzest/query/optimizer/tasks.py,sha256=pPDIV-efm_2S9WVYubu0QGuhh5-4c9SM3XtvfhpLDOY,24855
 palimpzest/query/processor/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 palimpzest/query/processor/config.py,sha256=dA3GB-nEAIQBllEx4rFy1W8iHYYEZOMOi4MgfZgODuQ,3851
@@ -84,12 +84,12 @@ palimpzest/utils/env_helpers.py,sha256=n81KzoJ459pRxo7QmJA7duazwWsfoMGTHc71D2Lat
 palimpzest/utils/field_helpers.py,sha256=Op18ThAnDlALiAkquUQbelHodZZYg378Ct1I8eIkKio,2291
 palimpzest/utils/generation_helpers.py,sha256=KTeVQDXRwM18auR8YsLNR5AdPDjSgO0NoOBagfFdTMs,3213
 palimpzest/utils/hash_helpers.py,sha256=3A8dA7SbXTwnnvZvPVNqqMLlVRhCKyKF_bjNNAu3Exk,334
-palimpzest/utils/model_helpers.py,sha256=utkjWvcKS7imF6Q39ef82pE8IRvDPoNPis1Mayp_N8w,3074
+palimpzest/utils/model_helpers.py,sha256=GEtWn0SIoTjEf94ZQZyP99zJ2sOEIYJYxhOCjaTcBO0,3433
 palimpzest/utils/progress.py,sha256=RiV753fKV8dFCNan4_wvdN2S3gvTvyRPq4Ksz6nSV1U,18536
 palimpzest/utils/sandbox.py,sha256=Ge96gmzqeOGlNkMCG9A95_PB8wRQbvTFua136of8FcA,6465
 palimpzest/utils/udfs.py,sha256=LjHic54B1az-rKgNLur0wOpaz2ko_UodjLEJrazkxvY,1854
-palimpzest-0.7.9.dist-info/licenses/LICENSE,sha256=5GUlHy9lr-Py9kvV38FF1m3yy3NqM18fefuE9wkWumo,1079
-palimpzest-0.7.9.dist-info/METADATA,sha256=FsXPsUUr-JN3jd6sPf5yGwpnNpCTCqXfNbvLmppSInU,7230
-palimpzest-0.7.9.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-palimpzest-0.7.9.dist-info/top_level.txt,sha256=raV06dJUgohefUn3ZyJS2uqp_Y76EOLA9Y2e_fxt8Ew,11
-palimpzest-0.7.9.dist-info/RECORD,,
+palimpzest-0.7.11.dist-info/licenses/LICENSE,sha256=5GUlHy9lr-Py9kvV38FF1m3yy3NqM18fefuE9wkWumo,1079
+palimpzest-0.7.11.dist-info/METADATA,sha256=dk4jW6VG9zMHNnzJBSUyRqzV1YPowR2lm-KG5wy7OdM,7231
+palimpzest-0.7.11.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+palimpzest-0.7.11.dist-info/top_level.txt,sha256=raV06dJUgohefUn3ZyJS2uqp_Y76EOLA9Y2e_fxt8Ew,11
+palimpzest-0.7.11.dist-info/RECORD,,