aryn-ai · bsowell · Oct 22, 2024 · Oct 21, 2024
diff --git a/lib/sycamore/sycamore/transforms/llm_query.py b/lib/sycamore/sycamore/transforms/llm_query.py
@@ -104,8 +104,17 @@ def _query_text_object(
                     .render(doc=object)
                 )
             else:
+                object_name = "ELEMENT" if isinstance(object, Element) else "DOCUMENT"
                 if objectPrev and objectPrev.text_representation:
-                    prompt = self._prompt + "\n" + objectPrev.text_representation + "\n\n" + object.text_representation
+                    prompt = (
+                        self._prompt
+                        + "\n"
+                        + f"{object_name} 1: \n\n"
+                        + objectPrev.text_representation
+                        + "\n\n"
+                        + f"{object_name} 2: \n"
+                        + object.text_representation
+                    )
                 else:
                     prompt = self._prompt + "\n" + object.text_representation
             prompt_kwargs = {"prompt": prompt}

diff --git a/lib/sycamore/sycamore/transforms/merge_elements.py b/lib/sycamore/sycamore/transforms/merge_elements.py
@@ -24,13 +24,11 @@ def should_merge(self, element1: Element, element2: Element) -> bool:
     def merge(self, element1: Element, element2: Element) -> Element:
         pass
 
-    @abstractmethod
     def preprocess_element(self, element: Element) -> Element:
-        pass
+        return element
 
-    @abstractmethod
     def postprocess_element(self, element: Element) -> Element:
-        pass
+        return element
 
     @timetrace("mergeElem")
     def merge_elements(self, document: Document) -> Document:
@@ -443,7 +441,7 @@ class TableMerger(ElementMerger):
             Respond with only 'true' or 'false' based on your certainty that the second table is a continuation. \
             Certainty is determined if either of the two conditions is true."
 
-            regex_pattern = r"table \d+"
+            regex_pattern = r"table \\d+"
 
             merger = TableMerger(llm_prompt = prompt, llm=llm)
 
@@ -481,8 +479,10 @@ def merge_elements(self, document: Document) -> Document:
         for element in table_elements[1:]:
             if self.should_merge(new_table_elements[-1], element):
                 new_table_elements[-1] = self.merge(new_table_elements[-1], element)
+                new_table_elements[-1]["properties"]["table_continuation"] = True
             else:
                 new_table_elements.append(element)
+                new_table_elements[-1]["properties"]["table_continuation"] = False
         other_elements.extend(new_table_elements)
         document.elements = other_elements
         bbox_sort_document(document)
@@ -607,12 +607,6 @@ def process_llm_query(self, document):
         llm_results = llm_query_agent.execute_query(document)
         return llm_results
 
-    def preprocess_element(self, elem: Element) -> Element:
-        return elem
-
-    def postprocess_element(self, elem: Element) -> Element:
-        return elem
-
 
 class Merge(SingleThreadUser, NonGPUUser, Map):
     """