my-aws-helpers 3.0.0.dev7__tar.gz → 3.1.0__tar.gz

{my_aws_helpers-3.0.0.dev7 → my_aws_helpers-3.1.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: my_aws_helpers
-Version: 3.0.0.dev7
+Version: 3.1.0
 Summary: AWS Helpers
 Home-page: https://github.com/JarrodMccarthy/aws_helpers.git
 Author: Jarrod McCarthy

{my_aws_helpers-3.0.0.dev7 → my_aws_helpers-3.1.0}/my_aws_helpers/bedrock.py

@@ -51,8 +51,9 @@ class OCRResult:
 class Bedrock:
     def __init__(
         self,
-        model_id: str = "anthropic.claude-3-5-sonnet-20241022-v2:0",
+        model_id: str = "apac.anthropic.claude-3-5-sonnet-20241022-v2:0",
         logger = None,
+        sleep_time: float = 1.0,
     ):
 
         self.session = Bedrock._set_session_params()
@@ -60,7 +61,7 @@ class Bedrock:
         region_name = "ap-southeast-2"
         if self.session is None:
             self.session = boto3.Session(region_name = region_name)
-
+        self.sleep_time = sleep_time
         
         custom_config = Config(
             retries={
@@ -133,6 +134,7 @@ class Bedrock:
             except Exception as e:
                 self.logger.exception(f"Error during conversation due to {e}")
                 if i >= len(retries) - 1: raise Exception(e)
+                time.sleep(self.sleep_time)
                 continue
         
         result = {}
@@ -144,24 +146,19 @@ class Bedrock:
     def _parallel_ocr(
             self, 
             image_bytes_list: List[bytes], 
-            prompt_type: str,
+            prompt: str,
             max_workers: int = 10,
     ):
-        results = list()
-        prompt = self._get_prompt(prompt_type=prompt_type)
+        execution_futures = []
         with concurrent.futures.ThreadPoolExecutor(max_workers=max_workers) as executor:
-            self.logger.info("Some log")
-            execution_futures = {
-                executor.submit(
-                    self._ocr, 
-                    prompt = prompt,
-                    image_bytes = img,
-                ): img for img in image_bytes_list
-            }
-            for future in concurrent.futures.as_completed(execution_futures):
-                result = future.result()
-                if result:
-                    results.append(result)
+            for i, img in enumerate(image_bytes_list):
+                self.logger.info(f"Starting OCR for page: {i}")
+                time.sleep(self.sleep_time)  # Stagger start time
+                future = executor.submit(self._ocr, prompt=prompt, image_bytes=img)
+                execution_futures.append(future)
+
+        # Wait for all tasks and collect results in order of submission
+        results = [future.result() for future in execution_futures if future.result() is not None]
         return results
     
     def get_ocr_result(
@@ -183,6 +180,7 @@ class Bedrock:
                 except Exception as e:
                     self.logger.error(f"Could not get pix map for page {i}")
                     continue
+            skip_page_zero = False
             header_ocr_result = None
             if len(image_bytes_list) > 1:
                 headers_prompt = self._get_prompt(prompt_type=PromptType.transaction_headers.value)
@@ -193,6 +191,7 @@ class Bedrock:
                         self.logger.info(f"No ocr result returned when getting headers {PromptType.transaction_headers.value}")
                     headers = header_ocr_result.content.get("headers")
                     if (len(headers) < 1) or (headers is None):
+                        skip_page_zero = True
                         continue
                     else:
                         break
@@ -204,9 +203,13 @@ class Bedrock:
             self.logger.info("Got Prompt")
             results = list()
 
-            for i, image_bytes in enumerate(image_bytes_list):
-                self.logger.info(f"Starting OCR for page: {i}")
-                results.append(self._ocr(image_bytes=image_bytes, prompt=transactions_prompt))
+            if skip_page_zero:
+                image_bytes_list = image_bytes_list[1:] #page zero often has account summary info
+            results = self._parallel_ocr(image_bytes_list=image_bytes_list, prompt=transactions_prompt)
+
+            # for i, image_bytes in enumerate(image_bytes_list):
+            #     self.logger.info(f"Starting OCR for page: {i}")
+            #     results.append(self._ocr(image_bytes=image_bytes, prompt=transactions_prompt))
             return results
         except Exception as e:
             self.logger.exception(e)

{my_aws_helpers-3.0.0.dev7 → my_aws_helpers-3.1.0}/my_aws_helpers/prompts/transactions_headers_prompt_v2.txt

@@ -13,7 +13,11 @@ Return a JSON object in the following structure:
 }
 
 ## Rules
-
+- DO NOT USE DETAILS sections, DO NOT use Account Summary, Do NOT use Account Details, DO NOT use payment details
+- ONLY Use tables that contain transaction to acquire headers, if you cannot find a table that contains a list of transactions, return:
+{
+  "headers": []
+}
 - Only return headers **if there is a table with transaction-like data** (e.g. purchases, sales, account activity).
 - There will likely be MORE THAN 1 transaction in this table name, if that is not the case, return:
 {

{my_aws_helpers-3.0.0.dev7 → my_aws_helpers-3.1.0}/my_aws_helpers.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: my-aws-helpers
-Version: 3.0.0.dev7
+Version: 3.1.0
 Summary: AWS Helpers
 Home-page: https://github.com/JarrodMccarthy/aws_helpers.git
 Author: Jarrod McCarthy

{my_aws_helpers-3.0.0.dev7 → my_aws_helpers-3.1.0}/setup.py

@@ -3,7 +3,7 @@ from setuptools import find_namespace_packages, setup
 
 base_path = os.path.abspath(os.path.dirname(__file__))
 
-version = "3.0.0.dev7"
+version = "3.1.0"
 
 setup(
     name="my_aws_helpers",