initial support for ISIN, AMFI code mapping for schemes.

Author: codereverser

casparser/process/cas_detailed.py

@@ -9,6 +9,7 @@
 from .regex import DETAILED_DATE_RE, FOLIO_RE, SCHEME_RE, REGISTRAR_RE
 from .regex import CLOSE_UNITS_RE, NAV_RE, OPEN_UNITS_RE, VALUATION_RE
 from .regex import DESCRIPTION_TAIL_RE, DIVIDEND_RE, TRANSACTION_RE
+from .utils import isin_search
 
 
 def parse_header(text):
@@ -86,7 +87,7 @@ def process_detailed_text(text):
         # "Registrar" column to the previous line
         if re.search(REGISTRAR_RE, line):
             line = "\t\t".join([lines[idx + 1], line])
-        elif m := re.search(DESCRIPTION_TAIL_RE, line, re.I | re.DOTALL):
+        if m := re.search(DESCRIPTION_TAIL_RE, line, re.I | re.DOTALL):
             description_tail = m.group(1).rstrip()
             line = line.replace(description_tail, "")
         else:
@@ -118,17 +119,21 @@ def process_detailed_text(text):
                 advisor = m.group(3)
                 if advisor is not None:
                     advisor = advisor.strip()
+                rta = m.group(4).strip()
+                rta_code = m.group(1).strip()
+                isin, amfi = isin_search(scheme, rta, rta_code)
                 curr_scheme_data = {
                     "scheme": scheme,
                     "advisor": advisor,
-                    "rta_code": m.group(1).strip(),
-                    "rta": m.group(4).strip(),
+                    "rta_code": rta_code,
+                    "rta": rta,
+                    "isin": isin,
+                    "amfi": amfi,
                     "open": Decimal(0.0),
                     "close": Decimal(0.0),
                     "valuation": {"date": None, "value": 0, "nav": 0},
                     "transactions": [],
                 }
-                balance = Decimal(0.0)
         if not curr_scheme_data:
             continue
         if m := re.search(OPEN_UNITS_RE, line):

casparser/process/cas_summary.py

@@ -6,6 +6,7 @@
 from ..enums import CASFileType
 from ..exceptions import HeaderParseError
 from .regex import SUMMARY_DATE_RE, SUMMARY_ROW_RE
+from .utils import isin_search
 
 
 def parse_header(text):
@@ -51,11 +52,16 @@ def process_summary_text(text):
             if curr_scheme_data.get("scheme") != scheme:
                 if curr_scheme_data:
                     folios[current_folio]["schemes"].append(curr_scheme_data)
+                rta = m.group(8).strip()
+                rta_code = m.group(2).strip()
+                isin, amfi = isin_search(scheme, rta, rta_code)
                 curr_scheme_data = {
                     "scheme": scheme,
                     "advisor": "N/A",
-                    "rta_code": m.group(2).strip(),
-                    "rta": m.group(8).strip(),
+                    "rta_code": rta_code,
+                    "rta": rta,
+                    "isin": isin,
+                    "amfi": amfi,
                     "open": Decimal(m.group(4).replace(",", "_")),
                     "close": Decimal(m.group(4).replace(",", "_")),
                     "valuation": {

casparser/process/utils.py

@@ -0,0 +1,19 @@
+from typing import Optional, Tuple
+
+from casparser_isin import MFISINDb
+
+
+def isin_search(scheme_name: str, rta: str, rta_code: str) -> Tuple[Optional[str], Optional[str]]:
+    """
+    Search isin db for ISIN and AMFI code
+
+    :param scheme_name: Scheme name from CAS
+    :param rta: RTA for the scheme
+    :param rta_code: Scheme RTA code
+    """
+    try:
+        with MFISINDb() as db:
+            scheme_data = db.isin_lookup(scheme_name, rta, rta_code)
+            return scheme_data.isin, scheme_data.amfi_code
+    except ValueError:
+        return None, None

casparser/types.py

@@ -1,6 +1,6 @@
 from datetime import date
 from decimal import Decimal
-from typing import List, TypedDict, Union
+from typing import Optional, List, TypedDict, Union
 
 from .enums import FileType, TransactionType
 
@@ -45,6 +45,8 @@ class SchemeType(TypedDict, total=False):
     advisor: str
     rta_code: str
     rta: str
+    isin: Optional[str]
+    amfi: Optional[str]
     open: Union[Decimal, float]
     close: Union[Decimal, float]
     valuation: SchemeValuationType

requirements.txt

@@ -1,3 +1,4 @@
+casparser-isin==0.1.0
 click==7.1.2
 colorama==0.4.4
 pdfminer.six==20201018

setup.cfg

@@ -21,6 +21,7 @@ include_package_data = True
 packages = find:
 python_requires = >=3.8
 install_requires =
+    casparser==0.1.0
     click==7.1.2
     colorama==0.4.4
     pdfminer.six==20201018