Added textract ocr parquet file and a limited README with new column definitions

README.md

@@ -1,3 +1,30 @@
 ---
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: chronam_ocr_1770-1810.with_ocr_text_new.parquet 
 license: cc0-1.0
 ---
+
+# Textract OCR data for the Chronicling America: Historic American Newspapers 1770–1810 dataset
+
+Column descriptions and definitions (and more important documentation) can be found in the source dataset at https://huggingface.co/datasets/RevolutionCrossroads/loc_chronicling_america_1770-1810.
+
+## Dataset Structure
+
+**Record Level**  
+Each record in the dataset corresponds to a single newspaper page. Issue-level metadata (title, place, date, edition) is repeated across all pages of that issue.
+
+## Data Fields
+
+New columns in this derivative dataset are:
+
+- **textract_OCR** *(string)*  
+OCR results from the AWS Textract API 
+
+- **textract_date** *(datetime64)*  
+The date and time that the Textract OCR was returned.
+
+- **textract_source** *(string)*  
+The source for the OCR data.

chronam_ocr_1770-1810.with_ocr_text_new.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0dda4a1e83bfc2aa3fb9651062a95fcf35824ddd291da03bdc3cb09a395c7e47
+size 528110874