BudSpender | Cannabis Receipt Parser
Your cannabis receipts hold a lot of data, put that data to work with BudSpender. This OpenAI-powered tool parses cannabis receipts into well-structured data that can be used for trending, analytics, and however you please.
Usage
Initialize a ReceiptsParser parsing client:
from cannlytics.data.sales import ReceiptsParser
# Initialize a COA parser.
parser = ReceiptsParser(openai_api_key=openai_api_key)
Parse a cannabis receipt:
# Parse a receipt.
filename = 'receipt-2023-04-20.jpeg'
data = parser.parse(filename)
Extracted Data
BudSpender is designed to extract structured data from unstructured text commonly found in cannabis receipts. Below is a breakdown of the extracted fields
| Field | Example | Description |
|---|---|---|
date_sold |
"2020-04-20" | The date the receipt was sold. |
invoice_number |
"123456789" | The receipt number. |
product_names |
["Blue Rhino Pre-Roll"] | The names of the product purchased. |
strain_names |
["Blue Rhino"] | The strain names of the products purchased. |
product_types |
["flower"] | The types of the products purchased. |
product_quantities |
[1] | The quantities of the products purchased. |
product_weights |
[3.5g] | The weights of the products purchased. |
product_prices |
[5.0] | The prices of the products purchased. |
product_ids |
["5f8b9c4b0f5c4b0008d1b2b0"] | The IDs of the products purchased. |
total_amount |
5.0 | The total amount of all product prices. |
subtotal |
5.0 | The subtotal of the receipt. |
total_discount |
0.0 | The amount of discount applied to the transaction, if applicable. |
total_paid |
5.0 | The total amount paid. |
change_due |
0.0 | The amount of change due. |
rewards_earned |
0.0 | The amount of rewards earned. |
rewards_spent |
0.0 | The amount of rewards spent. |
total_rewards |
0.0 | The total amount of rewards. |
city_tax |
0.0 | The amount of city tax applied to the transaction, if applicable. |
county_tax |
0.0 | The amount of county tax applied to the transaction, if applicable. |
state_tax |
0.0 | The amount of state tax applied to the transaction, if applicable. |
excise_tax |
0.0 | The amount of excise tax applied to the transaction, if applicable. |
retailer |
"BudHouse" | The name of the retailer. |
retailer_license_number |
"C11-0000001-LIC" | The license number of the retailer. |
retailer_address |
"1234 Main St, San Diego, CA 92101" | The address of the retailer. |
retailer_street |
"420 State Ave" | The retailer street, if applicable. |
retailer_city |
"Olympia" | The retailer city, if applicable. |
retailer_state |
"CA" | The state of the retailer, if applicable. |
retailer_zipcode |
"98506" | The zip code of the retailer, if applicable. |
budtender |
"John Doe" | The name of the budtender. |
Methods
| Method Name | Return Type | Description | Parameters |
|---|---|---|---|
parse |
dict |
Parses a receipt with OpenAI's GPT model and returns the structured data as a JSON. | doc: Path to the document.Other parameters to fine-tune the parsing such as model, openai_api_key, max_tokens, etc. |
image_to_pdf_to_text |
str |
Extracts the text from an image by converting it to a PDF. | image_file: Path to the image file. |
image_to_text |
str |
Directly extracts the text from an image. | image_file: Path to the image file.median_blur: (Optional) Removes noise. Must be a positive odd integer. |
save |
- | Saves the extracted data to a specified file. | obs: The parsed observation.filename: Destination file name. |
quit |
- | Resets the parser and performs garbage cleaning. | - |