diff options
Diffstat (limited to 'python')
| -rw-r--r-- | python/quote_parsing/parse_emails.py | 43 |
1 files changed, 40 insertions, 3 deletions
diff --git a/python/quote_parsing/parse_emails.py b/python/quote_parsing/parse_emails.py index 3e51b551..96d82115 100644 --- a/python/quote_parsing/parse_emails.py +++ b/python/quote_parsing/parse_emails.py @@ -134,6 +134,17 @@ def parse_refline(line): return d +def parse_refline_eu(line): + regex = r"Expiry:\s(?P<expiry>\S+)\s+Fwd:\s+(?P<fwd>\S+)\s+Fwd RBPV:\s+(?P<fwd_rbpv>\S+)" + m = re.match(regex, line) + if m := re.match(regex, line): + d = m.groupdict() + d["expiry"] = pd.to_datetime(d["expiry"], format="%d-%b-%Y") + else: + raise RuntimeError(f"can't parse refline {line}") + return d + + def parse_baml(fh, index_desc, *args): option_stack = {} fwd_index = [] @@ -157,13 +168,39 @@ def parse_baml(fh, index_desc, *args): return option_stack, fwd_index +def parse_baml_eu(fh, index_desc, *args): + option_stack = {} + fwd_index = [] + line = "" + while True: + if line == "": + try: + line = next(fh) + except StopIteration: + break + if line.startswith("BofA"): + df, line = parse_baml_block(fh, index_desc["index"]) + if line == "": + line = next(fh) + if line.startswith("Expiry"): + d = parse_refline_eu(line) + d.update(index_desc) + option_stack[ + (d["expiry"], index_desc["series"], index_desc["version"]) + ] = df + fwd_index.append(d) + else: + line = "" + return option_stack, fwd_index + + def parse_baml_block(fh, indextype): next(fh) # skip header r = [] line = "" for line in fh: line = line.strip() - if line.startswith("Ref") or line == "": + if line.startswith("Ref") or line == "" or line.startswith("Expiry"): break line = re.sub("[/|]", " ", line) vals = re.sub(" +", " ", line).rstrip().split(" ") @@ -801,10 +838,10 @@ regex_dict = { ), re.compile(r"(?:Fwd:){0,2}(?:BofA )?(XOVER|MAIN) S(\d{1,2}).*Ref[^\d]*([\d.]+)"): ( "BAML", - parse_baml, + parse_baml_eu, ), re.compile( - r"[^$]*(?:\$|\€){1,2} MS (?:CDX|ITX) OPTIONS: (IG|HY|MAIN|XO|FIN SEN)(\d{2})[^-]*- REF[^\d]*([\d.]+)" + r"[^$]*(?:\$|\€){1,2} ? MS (?:CDX|ITX) OPTIONS: (IG|HY|MAIN|XO|FIN SEN)(\d{2})[^-]*- REF[^\d]*([\d.]+)" ): ("MS", parse_ms), re.compile(r"(?:Fwd:)?CDX (IG|HY)(\d{2}).*- REF:[^\d]*([\d.]+)"): ( "NOM", |