palimpzest 0.8.5__tar.gz → 0.8.7__tar.gz

{palimpzest-0.8.5/src/palimpzest.egg-info → palimpzest-0.8.7}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: palimpzest
-Version: 0.8.5
+Version: 0.8.7
 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-0.8.5 → palimpzest-0.8.7}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "palimpzest"
-version = "0.8.5"
+version = "0.8.7"
 description = "Palimpzest is a system which enables anyone to process AI-powered analytical queries simply by defining them in a declarative language"
 readme = "README.md"
 requires-python = ">=3.10"

{palimpzest-0.8.5 → palimpzest-0.8.7}/src/palimpzest/core/elements/records.py

@@ -11,6 +11,8 @@ from pydantic.fields import FieldInfo
 
 from palimpzest.core.data import context
 from palimpzest.core.lib.schemas import (
+    AUDIO_FIELD_TYPES,
+    IMAGE_FIELD_TYPES,
     AudioBase64,
     AudioFilepath,
     ImageBase64,
@@ -303,9 +305,11 @@ class DataRecord:
             dct = {k: v for k, v in dct.items() if k in project_field_names}
 
         if not include_bytes:
+            bytes_field_types = [bytes, list[bytes], bytes | None, list[bytes] | None, bytes | Any, list[bytes] | Any]
+            bytes_field_types += AUDIO_FIELD_TYPES + IMAGE_FIELD_TYPES
             for k in dct:
                 field_type = self.get_field_type(k)
-                if field_type.annotation in [bytes, AudioBase64, ImageBase64, list[bytes], list[ImageBase64]]:
+                if field_type.annotation in bytes_field_types:
                     dct[k] = "<bytes>"
 
         if bytes_to_str:

{palimpzest-0.8.5 → palimpzest-0.8.7}/src/palimpzest/core/lib/schemas.py

@@ -33,20 +33,27 @@ IMAGE_LIST_FIELD_TYPES = [
     list[ImageBase64] | None,
     list[ImageFilepath] | None,
     list[ImageURL] | None,
+    list[ImageBase64] | Any,
+    list[ImageFilepath] | Any,
+    list[ImageURL] | Any,
 ]
 IMAGE_FIELD_TYPES = IMAGE_LIST_FIELD_TYPES + [
     ImageBase64, ImageFilepath, ImageURL,
     ImageBase64 | None, ImageFilepath | None, ImageURL | None,
+    ImageBase64 | Any, ImageFilepath | Any, ImageURL | Any,
 ]
 AUDIO_LIST_FIELD_TYPES = [
     list[AudioBase64],
     list[AudioFilepath],
     list[AudioBase64] | None,
     list[AudioFilepath] | None,
+    list[AudioBase64] | Any,
+    list[AudioFilepath] | Any,
 ]
 AUDIO_FIELD_TYPES = AUDIO_LIST_FIELD_TYPES + [
     AudioBase64, AudioFilepath,
     AudioBase64 | None, AudioFilepath | None,
+    AudioBase64 | Any, AudioFilepath | Any,
 ]
 
 

{palimpzest-0.8.5 → palimpzest-0.8.7}/src/palimpzest/prompts/prompt_factory.py

@@ -2,6 +2,7 @@
 
 import base64
 import json
+from typing import Any
 
 from pydantic import BaseModel
 
@@ -747,14 +748,14 @@ class PromptFactory:
             field_type = candidate.get_field_type(field_name)
 
             # audio filepath (or list of audio filepaths)
-            if field_type.annotation in [AudioFilepath, AudioFilepath | None]:
+            if field_type.annotation in [AudioFilepath, AudioFilepath | None, AudioFilepath | Any]:
                 with open(field_value, "rb") as f:
                     base64_audio_str = base64.b64encode(f.read()).decode("utf-8")
                 audio_content.append(
                     {"type": "input_audio", "input_audio": {"data": base64_audio_str, "format": "wav"}}
                 )
 
-            elif field_type.annotation in [list[AudioFilepath], list[AudioFilepath] | None]:
+            elif field_type.annotation in [list[AudioFilepath], list[AudioFilepath] | None, list[AudioFilepath] | Any]:
                 for audio_filepath in field_value:
                     with open(audio_filepath, "rb") as f:
                         base64_audio_str = base64.b64encode(f.read()).decode("utf-8")
@@ -763,12 +764,12 @@ class PromptFactory:
                     )
 
             # pre-encoded images (or list of pre-encoded images)
-            elif field_type.annotation in [AudioBase64, AudioBase64 | None]:
+            elif field_type.annotation in [AudioBase64, AudioBase64 | None, AudioBase64 | Any]:
                 audio_content.append(
                     {"type": "input_audio", "input_audio": {"data": field_value, "format": "wav"}}
                 )
 
-            elif field_type.annotation in [list[AudioBase64], list[AudioBase64] | None]:
+            elif field_type.annotation in [list[AudioBase64], list[AudioBase64] | None, list[AudioBase64] | Any]:
                 for base64_audio in field_value:
                     audio_content.append(
                         {"type": "input_audio", "input_audio": {"data": base64_audio, "format": "wav"}}
@@ -794,14 +795,14 @@ class PromptFactory:
             field_type = candidate.get_field_type(field_name)
 
             # image filepath (or list of image filepaths)
-            if field_type.annotation in [ImageFilepath, ImageFilepath | None]:
+            if field_type.annotation in [ImageFilepath, ImageFilepath | None, ImageFilepath | Any]:
                 with open(field_value, "rb") as f:
                     base64_image_str = base64.b64encode(f.read()).decode("utf-8")
                 image_content.append(
                     {"type": "image_url", "image_url": {"url": f"data:image/jpeg;base64,{base64_image_str}"}}
                 )
 
-            elif field_type.annotation in [list[ImageFilepath], list[ImageFilepath] | None]:
+            elif field_type.annotation in [list[ImageFilepath], list[ImageFilepath] | None, list[ImageFilepath] | Any]:
                 for image_filepath in field_value:
                     with open(image_filepath, "rb") as f:
                         base64_image_str = base64.b64encode(f.read()).decode("utf-8")
@@ -810,20 +811,20 @@ class PromptFactory:
                     )
 
             # image url (or list of image urls)
-            elif field_type.annotation in [ImageURL, ImageURL | None]:
+            elif field_type.annotation in [ImageURL, ImageURL | None, ImageURL | Any]:
                 image_content.append({"type": "image_url", "image_url": {"url": field_value}})
 
-            elif field_type.annotation in [list[ImageURL], list[ImageURL] | None]:
+            elif field_type.annotation in [list[ImageURL], list[ImageURL] | None, list[ImageURL] | Any]:
                 for image_url in field_value:
                     image_content.append({"type": "image_url", "image_url": {"url": image_url}})
 
             # pre-encoded images (or list of pre-encoded images)
-            elif field_type.annotation in [ImageBase64, ImageBase64 | None]:
+            elif field_type.annotation in [ImageBase64, ImageBase64 | None, ImageBase64 | Any]:
                 image_content.append(
                     {"type": "image_url", "image_url": {"url": f"data:image/jpeg;base64,{field_value}"}}
                 )
 
-            elif field_type.annotation in [list[ImageBase64], list[ImageBase64] | None]:
+            elif field_type.annotation in [list[ImageBase64], list[ImageBase64] | None, list[ImageBase64] | Any]:
                 for base64_image in field_value:
                     image_content.append(
                         {"type": "image_url", "image_url": {"url": f"data:image/jpeg;base64,{base64_image}"}}

{palimpzest-0.8.5 → palimpzest-0.8.7}/src/palimpzest/query/execution/mab_execution_strategy.py

@@ -755,7 +755,6 @@ class MABExecutionStrategy(SentinelExecutionStrategy):
 
     def execute_sentinel_plan(self, plan: SentinelPlan, train_dataset: dict[str, Dataset], validator: Validator) -> SentinelPlanStats:
         logger.info(f"Executing plan {plan.plan_id} with {self.max_workers} workers")
-        logger.info(f"Plan Details: {plan}")
 
         # initialize plan stats
         plan_stats = SentinelPlanStats.from_plan(plan)

{palimpzest-0.8.5 → palimpzest-0.8.7}/src/palimpzest/query/optimizer/optimizer_strategy.py

@@ -58,7 +58,6 @@ class GreedyStrategy(OptimizationStrategy):
     def get_optimal_plans(self, groups: dict, final_group_id: int, policy: Policy, use_final_op_quality: bool) -> list[PhysicalPlan]:
         logger.info(f"Getting greedy optimal plans for final group id: {final_group_id}")
         plans = [self._get_greedy_physical_plan(groups, final_group_id)]
-        logger.info(f"Greedy optimal plans: {plans}")
         logger.info(f"Done getting greedy optimal plans for final group id: {final_group_id}")
 
         return plans
@@ -137,7 +136,6 @@ class ParetoStrategy(OptimizationStrategy):
             optimal_plan = optimal_plan if policy.choose(optimal_plan.plan_cost, plan.plan_cost) else plan
 
         plans = [optimal_plan]
-        logger.info(f"Pareto optimal plans: {plans}")
         logger.info(f"Done getting pareto optimal plans for final group id: {final_group_id}")
         return plans
     
@@ -174,7 +172,6 @@ class SentinelStrategy(OptimizationStrategy):
     def get_optimal_plans(self, groups: dict, final_group_id: int, policy: Policy, use_final_op_quality: bool) -> list[SentinelPlan]:
         logger.info(f"Getting sentinel optimal plans for final group id: {final_group_id}")
         plans = [self._get_sentinel_plan(groups, final_group_id)]
-        logger.info(f"Sentinel optimal plans: {plans}")
         logger.info(f"Done getting sentinel optimal plans for final group id: {final_group_id}")
         return plans
 

{palimpzest-0.8.5 → palimpzest-0.8.7}/src/palimpzest/query/optimizer/plan.py

@@ -330,12 +330,11 @@ class SentinelPlan(Plan):
 
     def _get_str(self, idx: int = 0, indent: int = 0) -> str:
         indent_str = " " * (indent * 2)
-        plan_str = ""
-        for inner_idx, operator in enumerate(self.operator_set):
-            inner_idx_str = "" if len(self.operator_set) == 1 else f"{inner_idx + 1}."
-            plan_str += f"{indent_str}{idx}.{inner_idx_str} {str(operator)}\n"
-            for subplan in self.subplans:
-                plan_str += subplan._get_str(idx=idx + 1, indent=indent + 1)
+        operator = self.operator_set[0]
+        inner_idx_str = "" if len(self.operator_set) == 1 else f"1 - {len(self.operator_set)}."
+        plan_str = f"{indent_str}{idx}.{inner_idx_str} {str(operator)}\n"
+        for subplan in self.subplans:
+            plan_str += subplan._get_str(idx=idx + 1, indent=indent + 1)
 
         return plan_str
 

{palimpzest-0.8.5 → palimpzest-0.8.7}/src/palimpzest/validator/validator.py

@@ -2,8 +2,8 @@ import json
 import time
 
 import litellm
-from colorama import Fore, Style
 
+# from colorama import Fore, Style
 from palimpzest.constants import MODEL_CARDS, Cardinality, Model, PromptStrategy
 from palimpzest.core.elements.records import DataRecord
 from palimpzest.core.models import GenerationStats
@@ -87,7 +87,7 @@ class Validator:
         input_messages = [msg for msg in messages if msg["role"] != "system"]
         output = json.dumps(output, indent=2)
         output_message = f"OUTPUT:\n--------\n{output}\n\nEVALUATION: "
-        input_str = '\n'.join(list(map(lambda d: d['content'], input_messages + [{"role": "user", "content": output_message}])))
+        # input_str = '\n'.join(list(map(lambda d: d['content'], input_messages + [{"role": "user", "content": output_message}])))
 
         # invoke the judge
         score, gen_stats = None, GenerationStats()
@@ -98,8 +98,8 @@ class Validator:
             completion = litellm.completion(model=self.model.value, messages=val_messages)
             completion_text = completion.choices[0].message.content
             gen_stats = self._get_gen_stats_from_completion(completion, start_time)
-            print(f"INPUT:\n{input_str}")
-            print(Fore.GREEN + f"{completion_text}\n" + Style.RESET_ALL)
+            # print(f"INPUT:\n{input_str}")
+            # print(Fore.GREEN + f"{completion_text}\n" + Style.RESET_ALL)
 
             # parse the evaluation
             eval_dict: dict = get_json_from_answer(completion_text, self.model, Cardinality.ONE_TO_ONE)
@@ -233,7 +233,7 @@ class Validator:
         input_messages = [msg for msg in messages if msg["role"] != "system"]
         output = json.dumps(output, indent=2)
         output_message = f"OUTPUT:\n--------\n{output}\n\nEVALUATION: "
-        input_str = '\n'.join(list(map(lambda d: d['content'], input_messages + [{"role": "user", "content": output_message}])))
+        # input_str = '\n'.join(list(map(lambda d: d['content'], input_messages + [{"role": "user", "content": output_message}])))
 
         # invoke the judge
         score, gen_stats = None, GenerationStats()
@@ -245,8 +245,8 @@ class Validator:
             completion = litellm.completion(model="openai/o4-mini", messages=val_messages)
             completion_text = completion.choices[0].message.content
             gen_stats = self._get_gen_stats_from_completion(completion, start_time)
-            print(f"INPUT:\n{input_str}")
-            print(Fore.GREEN + f"{completion_text}\n" + Style.RESET_ALL)
+            # print(f"INPUT:\n{input_str}")
+            # print(Fore.GREEN + f"{completion_text}\n" + Style.RESET_ALL)
 
             # parse the evaluation
             eval_dict: dict = get_json_from_answer(completion_text, self.model, Cardinality.ONE_TO_ONE)

{palimpzest-0.8.5 → palimpzest-0.8.7/src/palimpzest.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: palimpzest
-Version: 0.8.5
+Version: 0.8.7
 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