bolade/ds_quickbooks
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Add new fields for tax and depreciation in receipt processing

Browse Source 
- Introduced fields: receipt_location, calculated_tax, is_depreciable, cca_rate, useful_life, and residual_value in DBReceipt model.
- Updated process_document function to handle new receipt data attributes.
- Enhanced DocumentProcessResponse schema to include new fields.
- Updated document processing rules to incorporate tax calculation based on location and depreciation rules.
This commit is contained in:
bolade
2025-10-07 11:15:26 +01:00
parent 823c05f78d
commit d8315f13ac
5 changed files with 172 additions and 23 deletions
Download Patch File Download Diff File Expand all files Collapse all files
app/services/document_processor.py
+137 -12
View File
@@ -7,7 +7,6 @@ from typing import Any, Dict
import aiofiles
import groq
import PyPDF2
from config import settings
logger = logging.getLogger(__name__)
@@ -47,18 +46,55 @@ class DocumentProcessor:
"date": "YYYY-MM-DD",
"category": "Food/Transport/Office/Other",
"confidence": 0.95,
"currency": "USD"
"currency": "USD",
"location": "Province/State, Country",
"calculated_tax": 0.00,
"is_depreciable": false,
"cca_rate": null,
"useful_life": null,
"residual_value": null
}
Rules:
- Extract vendor name as it appears on receipt
- Extract description of items/services purchased (e.g., "Coffee and sandwich", "Gasoline", "Office supplies")
- Total amount should be the final total including tax
- Tax amount is separate tax line if available
- Tax amount is separate tax line if available (if not clearly shown, calculate based on location)
- Date should be the date on the receipt
- Categorize based on vendor type (Starbucks=Food, Shell=Transport, etc.)
- Confidence score 0-1 based on how clear the receipt is
- Currency should be the currency used on the receipt (e.g., "USD", "EUR")
- Currency should be the currency used on the receipt (e.g., "USD", "EUR", "CAD")
LOCATION & TAX RULES:
- Extract location from receipt (look for store address, province/state, country)
- Format location as "Province/State, Country" (e.g., "Ontario, Canada" or "California, USA")
- If location not shown, return null for location
- If tax_amount is NOT clearly shown on receipt, calculate it based on location:
* Ontario, Canada: 13% HST
* Quebec, Canada: 9.975% QST + 5% GST = 14.975% total
* British Columbia, Canada: 12% (5% GST + 7% PST)
* Alberta, Canada: 5% GST
* California, USA: ~7.25% (varies by locality)
* New York, USA: ~8.875% (varies by locality)
* Texas, USA: 6.25%
* For other locations, estimate based on typical rates
- Store calculated tax in "calculated_tax" field (set to null if tax clearly shown)
- If tax is clearly shown on receipt, use that value for tax_amount and set calculated_tax to null
DEPRECIATION RULES:
- Determine if item is a depreciable asset (vehicles, machinery, equipment, computers, furniture, buildings)
- Set is_depreciable to true only for capital assets, false for consumables/services
- If is_depreciable is true, provide:
* cca_rate: CCA rate as decimal (e.g., 0.30 for 30%, 0.20 for 20%, 0.04 for 4%)
- Class 10 (Vehicles): 30%
- Class 8 (Furniture, equipment): 20%
- Class 50 (Computers, software): 55%
- Class 1 (Buildings): 4%
- Class 10.1 (Passenger vehicles >$30k): 30%
* useful_life: Expected years of use (e.g., 5 for computers, 8 for vehicles, 10 for furniture)
* residual_value: Estimated value at end of life (typically 10% of purchase price for equipment, 20% for vehicles)
- If is_depreciable is false, set cca_rate, useful_life, and residual_value to null
Return only valid JSON.
"""
@@ -80,7 +116,7 @@ class DocumentProcessor:
}
],
model=self.model,
max_tokens=500,
max_tokens=800,
temperature=0.1,
)
@@ -136,18 +172,55 @@ class DocumentProcessor:
"date": "YYYY-MM-DD",
"category": "Food/Transport/Office/Other",
"confidence": 0.95,
"currency": "USD"
"currency": "USD",
"location": "Province/State, Country",
"calculated_tax": 0.00,
"is_depreciable": false,
"cca_rate": null,
"useful_life": null,
"residual_value": null
}}
Rules:
- Extract vendor name as it appears on receipt
- Extract description of items/services purchased (e.g., "Coffee and sandwich", "Gasoline", "Office supplies")
- Total amount should be the final total including tax
- Tax amount is separate tax line if available
- Tax amount is separate tax line if available (if not clearly shown, calculate based on location)
- Date should be the date on the receipt
- Categorize based on vendor type
- Confidence score 0-1 based on clarity
- Currency should be the currency used on the receipt (e.g., "USD", "EUR")
- Currency should be the currency used on the receipt (e.g., "USD", "EUR", "CAD")
LOCATION & TAX RULES:
- Extract location from receipt (look for store address, province/state, country)
- Format location as "Province/State, Country" (e.g., "Ontario, Canada" or "California, USA")
- If location not shown, return null for location
- If tax_amount is NOT clearly shown on receipt, calculate it based on location:
* Ontario, Canada: 13% HST
* Quebec, Canada: 9.975% QST + 5% GST = 14.975% total
* British Columbia, Canada: 12% (5% GST + 7% PST)
* Alberta, Canada: 5% GST
* California, USA: ~7.25% (varies by locality)
* New York, USA: ~8.875% (varies by locality)
* Texas, USA: 6.25%
* For other locations, estimate based on typical rates
- Store calculated tax in "calculated_tax" field (set to null if tax clearly shown)
- If tax is clearly shown on receipt, use that value for tax_amount and set calculated_tax to null
DEPRECIATION RULES:
- Determine if item is a depreciable asset (vehicles, machinery, equipment, computers, furniture, buildings)
- Set is_depreciable to true only for capital assets, false for consumables/services
- If is_depreciable is true, provide:
* cca_rate: CCA rate as decimal (e.g., 0.30 for 30%, 0.20 for 20%, 0.04 for 4%)
- Class 10 (Vehicles): 30%
- Class 8 (Furniture, equipment): 20%
- Class 50 (Computers, software): 55%
- Class 1 (Buildings): 4%
- Class 10.1 (Passenger vehicles >$30k): 30%
* useful_life: Expected years of use (e.g., 5 for computers, 8 for vehicles, 10 for furniture)
* residual_value: Estimated value at end of life (typically 10% of purchase price for equipment, 20% for vehicles)
- If is_depreciable is false, set cca_rate, useful_life, and residual_value to null
Return only valid JSON.
"""
@@ -155,7 +228,7 @@ class DocumentProcessor:
response = self.client.chat.completions.create(
model=self.model,
messages=[{"role": "user", "content": prompt}],
max_tokens=500,
max_tokens=800,
temperature=0.1,
)
@@ -207,8 +280,22 @@ class DocumentProcessor:
confidence_match = re.search(
r'"confidence"\s*:\s*([0-9.]+)', json_str
)
currency_match = re.search(
r'"currency"\s*:\s*"([^"]*)"', json_str
currency_match = re.search(r'"currency"\s*:\s*"([^"]*)"', json_str)
location_match = re.search(r'"location"\s*:\s*"([^"]*)"', json_str)
calculated_tax_match = re.search(
r'"calculated_tax"\s*:\s*([0-9.]+|null)', json_str
)
is_depreciable_match = re.search(
r'"is_depreciable"\s*:\s*(true|false)', json_str
)
cca_rate_match = re.search(
r'"cca_rate"\s*:\s*([0-9.]+|null)', json_str
)
useful_life_match = re.search(
r'"useful_life"\s*:\s*([0-9]+|null)', json_str
)
residual_value_match = re.search(
r'"residual_value"\s*:\s*([0-9.]+|null)', json_str
)
data = {
@@ -229,7 +316,27 @@ class DocumentProcessor:
"confidence": float(confidence_match.group(1))
if confidence_match
else 0.5,
"currency": currency_match.group(1) if currency_match else "CAD"
"currency": currency_match.group(1)
if currency_match
else "CAD",
"location": location_match.group(1) if location_match else None,
"calculated_tax": float(calculated_tax_match.group(1))
if calculated_tax_match
and calculated_tax_match.group(1) != "null"
else None,
"is_depreciable": is_depreciable_match.group(1) == "true"
if is_depreciable_match
else None,
"cca_rate": float(cca_rate_match.group(1))
if cca_rate_match and cca_rate_match.group(1) != "null"
else None,
"useful_life": int(useful_life_match.group(1))
if useful_life_match and useful_life_match.group(1) != "null"
else None,
"residual_value": float(residual_value_match.group(1))
if residual_value_match
and residual_value_match.group(1) != "null"
else None,
}
# Validate and clean data
@@ -243,6 +350,12 @@ class DocumentProcessor:
"confidence": float(data.get("confidence", 0.5)),
"extraction_success": True,
"currency": data.get("currency", "CAD").strip(),
"location": data.get("location"),
"calculated_tax": data.get("calculated_tax"),
"is_depreciable": data.get("is_depreciable"),
"cca_rate": data.get("cca_rate"),
"useful_life": data.get("useful_life"),
"residual_value": data.get("residual_value"),
}
else:
# Try to extract fields from plain text
@@ -312,6 +425,12 @@ class DocumentProcessor:
"category": "Other",
"confidence": 0.3, # Low confidence for text extraction
"extraction_success": True,
"location": None,
"calculated_tax": None,
"is_depreciable": None,
"cca_rate": None,
"useful_life": None,
"residual_value": None,
}
except Exception as e:
@@ -325,6 +444,12 @@ class DocumentProcessor:
"confidence": 0.1,
"extraction_success": False,
"error": f"Text extraction failed: {str(e)}",
"location": None,
"calculated_tax": None,
"is_depreciable": None,
"cca_rate": None,
"useful_life": None,
"residual_value": None,
}
async def save_uploaded_file(self, file_content: bytes, filename: str) -> str: