Add mic_code (ISO 10383 MIC) column

[dokson]

Closes #110.

Adds a standardised ISO 10383 MIC code for each ticker, as requested in the issue, by mapping the existing Yahoo exchange code to its MIC.

Data

• New mic_code column (placed right after exchange) in equities, etfs, funds and indices.
• Every Yahoo exchange code that corresponds to a real trading venue is mapped to its operating MIC (e.g. NMS/NGM/NCM → XNAS, NYQ → XNYS, LSE/IOB → XLON, GER → XETR, JPX → XJPX).
• All mapped MICs were validated as ACTIVE against the official ISO 10383 registry, with the ISO country code cross-checked against the data.
• Pseudo-codes that are not trading venues (indices, FX, crypto, NAV mutual funds) and venue-ambiguous codes (e.g. ENX, generic Euronext) are intentionally left blank rather than guessed.
• Spot-checked end-to-end on well-known tickers (AAPL→XNAS, JPM→XNYS, VOD.L→XLON, SAP.DE→XETR, 7203.T→XJPX, NESN.SW→XSWX, 005930.KS→XKRX, …).

Package

• mic_code added as a filter to select() and as an option in show_options() for every asset class.
• The duplicated validate-and-filter blocks were refactored into a single shared helper (FinanceDatabase._filter_by_options) with a centralised COLUMN_LABELS map for the error messages.
• select() is now annotated to return FinanceFrame; search(**kwargs) typing broadened.
• Fixed a pandas DtypeWarning and cleared the package's pyright/ruff findings.

Tests

• New coverage for the mic_code filter, show_options, and an exchange → mic_code one-to-one invariant.
• Snapshot fixtures regenerated.

Note: compression/ artifacts are intentionally not committed — they are regenerated by the Database-Update workflow.

[JerBouma]

It says the column is right next to exchange, shouldn't it be alongside the other codes? E.g. after the ISIN column? Whatever makes the most sense.

Other columns are without the word "code" so I'd say we just make it "mic" (and MIC already contains the word code 😉)

EDIT: it's related to exchange so location of the column makes sense!

[dokson]

Good points, thanks!

On the name: agreed — I'll rename it to mic to stay consistent with the other columns (none use a code suffix).

On the placement: I put it right after exchange on purpose, because mic is essentially the ISO-standardised encoding of exchange — both identify the trading venue, just in different notations (the Yahoo code vs. ISO 10383). Keeping them adjacent (and next to market) lets you read the venue in both forms at a glance. isin / cusip / figi, on the other hand, identify the security itself rather than the venue, so semantically mic felt closer to exchange/market than to that block.

That said, it's your call — happy to move it next to isin with the other identifiers if you prefer that grouping. Just let me know and I'll update the PR (together with the mic rename).

[JerBouma]

Location makes sense indeed, if you could rename to mic this looks good to go!

[dokson]

Location makes sense indeed, if you could rename to mic this looks good to go!

renamed 😉

[JerBouma]

The code is making a lot of changes to the package, somewhat simplifying it but in this case how the code originally was written was done on purpose. This adds a layer of abstraction that, given the minimal size of the package, doesn't add much value.

I'd much prefer to keep the package the way it is even if it can in theory be done in less lines of code.

[dokson]

The code is making a lot of changes to the package, somewhat simplifying it but in this case how the code originally was written was done on purpose. This adds a layer of abstraction that, given the minimal size of the package, doesn't add much value.

I'd much prefer to keep the package the way it is even if it can in theory be done in less lines of code.

following your guide

[JerBouma]

Perfect, thank you. Once there is functionality for the delisted tickers (as mentioned in your issue) I'll release a new PyPi and GitHub release. Wanted to work on this somewhere next week.

[JerBouma]

@dokson Forgot something, the GitHub Actions script doesn't include a placeholder for the new mic column yet. It also should be defined based on new additions over there given the exchange code will be filled.

[dokson]

the GitHub Actions script doesn't include a placeholder for the new mic column yet. It also should be defined based on new additions over there given the exchange code will be filled.

@JerBouma fixed in #150. In the Database-Update action, build_new_ticker now derives exchange → ISO 10383 MIC from the existing one-to-one data and fills mic for new tickers (placed right after exchange); unknown codes intentionally stay blank. Since the job only ever assigns NMS/ASE, those resolve to XNAS/XASE.

It also backfills the 16 NASDAQ rows (NMS → XNAS) that were added by a prior run before the column existed, and adds test_mic_filled_when_exchange_mapped — which fails if a row has a known exchange but a missing mic (the gap that let those rows ship).