Merge pull request #366 from posit-dev/enh-dataset-generate-country-name-code

Enh: add dataset generation presets for country codes (2- and 3-letter variants)

Author: rich-iannone

docs/user-guide/test-data-generation.qmd

@@ -191,6 +191,8 @@ The `preset=` parameter in `string_field()` supports many data types:
 - `city`: city name
 - `state`: state/province name
 - `country`: country name
+- `country_code_2`: ISO 3166-1 alpha-2 country code (e.g., `"US"`)
+- `country_code_3`: ISO 3166-1 alpha-3 country code (e.g., `"USA"`)
 - `postcode`: postal/ZIP code
 - `latitude`: latitude coordinate
 - `longitude`: longitude coordinate

pointblank/countries/__init__.py

@@ -16,6 +16,7 @@
     "LocaleGenerator",
     "get_generator",
     "COUNTRY_CODE_MAP",
+    "COUNTRY_INFO",
     "COUNTRIES_WITH_FULL_DATA",
 ]
 
@@ -304,6 +305,252 @@
     "ZA",  # South Africa
 ]
 
+# Comprehensive country info mapping: alpha-2 -> (alpha-3, English name)
+COUNTRY_INFO: dict[str, tuple[str, str]] = {
+    "AD": ("AND", "Andorra"),
+    "AE": ("ARE", "United Arab Emirates"),
+    "AF": ("AFG", "Afghanistan"),
+    "AG": ("ATG", "Antigua & Barbuda"),
+    "AI": ("AIA", "Anguilla"),
+    "AL": ("ALB", "Albania"),
+    "AM": ("ARM", "Armenia"),
+    "AO": ("AGO", "Angola"),
+    "AR": ("ARG", "Argentina"),
+    "AS": ("ASM", "American Samoa"),
+    "AT": ("AUT", "Austria"),
+    "AU": ("AUS", "Australia"),
+    "AW": ("ABW", "Aruba"),
+    "AX": ("ALA", "\u00c5land Islands"),
+    "AZ": ("AZE", "Azerbaijan"),
+    "BA": ("BIH", "Bosnia & Herzegovina"),
+    "BB": ("BRB", "Barbados"),
+    "BD": ("BGD", "Bangladesh"),
+    "BE": ("BEL", "Belgium"),
+    "BF": ("BFA", "Burkina Faso"),
+    "BG": ("BGR", "Bulgaria"),
+    "BH": ("BHR", "Bahrain"),
+    "BI": ("BDI", "Burundi"),
+    "BJ": ("BEN", "Benin"),
+    "BL": ("BLM", "St. Barth\u00e9lemy"),
+    "BM": ("BMU", "Bermuda"),
+    "BN": ("BRN", "Brunei"),
+    "BO": ("BOL", "Bolivia"),
+    "BR": ("BRA", "Brazil"),
+    "BS": ("BHS", "Bahamas"),
+    "BT": ("BTN", "Bhutan"),
+    "BW": ("BWA", "Botswana"),
+    "BY": ("BLR", "Belarus"),
+    "BZ": ("BLZ", "Belize"),
+    "CA": ("CAN", "Canada"),
+    "CC": ("CCK", "Cocos (Keeling) Islands"),
+    "CD": ("COD", "Congo (DRC)"),
+    "CF": ("CAF", "Central African Republic"),
+    "CG": ("COG", "Congo (Republic)"),
+    "CH": ("CHE", "Switzerland"),
+    "CI": ("CIV", "C\u00f4te d\u2019Ivoire"),
+    "CK": ("COK", "Cook Islands"),
+    "CL": ("CHL", "Chile"),
+    "CM": ("CMR", "Cameroon"),
+    "CN": ("CHN", "China"),
+    "CO": ("COL", "Colombia"),
+    "CR": ("CRI", "Costa Rica"),
+    "CU": ("CUB", "Cuba"),
+    "CV": ("CPV", "Cape Verde"),
+    "CW": ("CUW", "Cura\u00e7ao"),
+    "CY": ("CYP", "Cyprus"),
+    "CZ": ("CZE", "Czech Republic"),
+    "DE": ("DEU", "Germany"),
+    "DJ": ("DJI", "Djibouti"),
+    "DK": ("DNK", "Denmark"),
+    "DM": ("DMA", "Dominica"),
+    "DO": ("DOM", "Dominican Republic"),
+    "DZ": ("DZA", "Algeria"),
+    "EC": ("ECU", "Ecuador"),
+    "EE": ("EST", "Estonia"),
+    "EG": ("EGY", "Egypt"),
+    "EH": ("ESH", "Western Sahara"),
+    "ER": ("ERI", "Eritrea"),
+    "ES": ("ESP", "Spain"),
+    "ET": ("ETH", "Ethiopia"),
+    "EU": ("EUR", "European Union"),
+    "FI": ("FIN", "Finland"),
+    "FJ": ("FJI", "Fiji"),
+    "FK": ("FLK", "Falkland Islands"),
+    "FM": ("FSM", "Micronesia"),
+    "FO": ("FRO", "Faroe Islands"),
+    "FR": ("FRA", "France"),
+    "GA": ("GAB", "Gabon"),
+    "GB": ("GBR", "United Kingdom"),
+    "GD": ("GRD", "Grenada"),
+    "GE": ("GEO", "Georgia"),
+    "GF": ("GUF", "French Guiana"),
+    "GG": ("GGY", "Guernsey"),
+    "GH": ("GHA", "Ghana"),
+    "GI": ("GIB", "Gibraltar"),
+    "GL": ("GRL", "Greenland"),
+    "GM": ("GMB", "Gambia"),
+    "GN": ("GIN", "Guinea"),
+    "GP": ("GLP", "Guadeloupe"),
+    "GQ": ("GNQ", "Equatorial Guinea"),
+    "GR": ("GRC", "Greece"),
+    "GS": ("SGS", "South Georgia & South Sandwich Islands"),
+    "GT": ("GTM", "Guatemala"),
+    "GU": ("GUM", "Guam"),
+    "GW": ("GNB", "Guinea-Bissau"),
+    "GY": ("GUY", "Guyana"),
+    "HK": ("HKG", "Hong Kong"),
+    "HN": ("HND", "Honduras"),
+    "HR": ("HRV", "Croatia"),
+    "HT": ("HTI", "Haiti"),
+    "HU": ("HUN", "Hungary"),
+    "ID": ("IDN", "Indonesia"),
+    "IE": ("IRL", "Ireland"),
+    "IL": ("ISR", "Israel"),
+    "IM": ("IMN", "Isle of Man"),
+    "IN": ("IND", "India"),
+    "IO": ("IOT", "British Indian Ocean Territory"),
+    "IQ": ("IRQ", "Iraq"),
+    "IR": ("IRN", "Iran"),
+    "IS": ("ISL", "Iceland"),
+    "IT": ("ITA", "Italy"),
+    "JE": ("JEY", "Jersey"),
+    "JM": ("JAM", "Jamaica"),
+    "JO": ("JOR", "Jordan"),
+    "JP": ("JPN", "Japan"),
+    "KE": ("KEN", "Kenya"),
+    "KG": ("KGZ", "Kyrgyzstan"),
+    "KH": ("KHM", "Cambodia"),
+    "KI": ("KIR", "Kiribati"),
+    "KM": ("COM", "Comoros"),
+    "KN": ("KNA", "St. Kitts & Nevis"),
+    "KP": ("PRK", "North Korea"),
+    "KR": ("KOR", "South Korea"),
+    "KW": ("KWT", "Kuwait"),
+    "KY": ("CYM", "Cayman Islands"),
+    "KZ": ("KAZ", "Kazakhstan"),
+    "LA": ("LAO", "Laos"),
+    "LB": ("LBN", "Lebanon"),
+    "LC": ("LCA", "St. Lucia"),
+    "LI": ("LIE", "Liechtenstein"),
+    "LK": ("LKA", "Sri Lanka"),
+    "LR": ("LBR", "Liberia"),
+    "LS": ("LSO", "Lesotho"),
+    "LT": ("LTU", "Lithuania"),
+    "LU": ("LUX", "Luxembourg"),
+    "LV": ("LVA", "Latvia"),
+    "LY": ("LBY", "Libya"),
+    "MA": ("MAR", "Morocco"),
+    "MC": ("MCO", "Monaco"),
+    "MD": ("MDA", "Moldova"),
+    "ME": ("MNE", "Montenegro"),
+    "MF": ("MAF", "St. Martin"),
+    "MG": ("MDG", "Madagascar"),
+    "MH": ("MHL", "Marshall Islands"),
+    "MK": ("MKD", "North Macedonia"),
+    "ML": ("MLI", "Mali"),
+    "MM": ("MMR", "Myanmar"),
+    "MN": ("MNG", "Mongolia"),
+    "MO": ("MAC", "Macao"),
+    "MP": ("MNP", "Northern Mariana Islands"),
+    "MQ": ("MTQ", "Martinique"),
+    "MR": ("MRT", "Mauritania"),
+    "MS": ("MSR", "Montserrat"),
+    "MT": ("MLT", "Malta"),
+    "MU": ("MUS", "Mauritius"),
+    "MV": ("MDV", "Maldives"),
+    "MW": ("MWI", "Malawi"),
+    "MX": ("MEX", "Mexico"),
+    "MY": ("MYS", "Malaysia"),
+    "MZ": ("MOZ", "Mozambique"),
+    "NA": ("NAM", "Namibia"),
+    "NC": ("NCL", "New Caledonia"),
+    "NE": ("NER", "Niger"),
+    "NF": ("NFK", "Norfolk Island"),
+    "NG": ("NGA", "Nigeria"),
+    "NI": ("NIC", "Nicaragua"),
+    "NL": ("NLD", "Netherlands"),
+    "NO": ("NOR", "Norway"),
+    "NP": ("NPL", "Nepal"),
+    "NR": ("NRU", "Nauru"),
+    "NU": ("NIU", "Niue"),
+    "NZ": ("NZL", "New Zealand"),
+    "OM": ("OMN", "Oman"),
+    "PA": ("PAN", "Panama"),
+    "PE": ("PER", "Peru"),
+    "PF": ("PYF", "French Polynesia"),
+    "PG": ("PNG", "Papua New Guinea"),
+    "PH": ("PHL", "Philippines"),
+    "PK": ("PAK", "Pakistan"),
+    "PL": ("POL", "Poland"),
+    "PM": ("SPM", "St. Pierre & Miquelon"),
+    "PN": ("PCN", "Pitcairn Islands"),
+    "PR": ("PRI", "Puerto Rico"),
+    "PS": ("PSE", "Palestine"),
+    "PT": ("PRT", "Portugal"),
+    "PW": ("PLW", "Palau"),
+    "PY": ("PRY", "Paraguay"),
+    "QA": ("QAT", "Qatar"),
+    "RE": ("REU", "R\u00e9union"),
+    "RO": ("ROU", "Romania"),
+    "RS": ("SRB", "Serbia"),
+    "RU": ("RUS", "Russia"),
+    "RW": ("RWA", "Rwanda"),
+    "SA": ("SAU", "Saudi Arabia"),
+    "SB": ("SLB", "Solomon Islands"),
+    "SC": ("SYC", "Seychelles"),
+    "SD": ("SDN", "Sudan"),
+    "SE": ("SWE", "Sweden"),
+    "SG": ("SGP", "Singapore"),
+    "SI": ("SVN", "Slovenia"),
+    "SK": ("SVK", "Slovakia"),
+    "SL": ("SLE", "Sierra Leone"),
+    "SM": ("SMR", "San Marino"),
+    "SN": ("SEN", "Senegal"),
+    "SO": ("SOM", "Somalia"),
+    "SR": ("SUR", "Suriname"),
+    "SS": ("SSD", "South Sudan"),
+    "ST": ("STP", "S\u00e3o Tom\u00e9 & Pr\u00edncipe"),
+    "SV": ("SLV", "El Salvador"),
+    "SX": ("SXM", "Sint Maarten"),
+    "SY": ("SYR", "Syria"),
+    "SZ": ("SWZ", "Eswatini"),
+    "TC": ("TCA", "Turks & Caicos Islands"),
+    "TD": ("TCD", "Chad"),
+    "TF": ("ATF", "French Southern Territories"),
+    "TG": ("TGO", "Togo"),
+    "TH": ("THA", "Thailand"),
+    "TJ": ("TJK", "Tajikistan"),
+    "TK": ("TKL", "Tokelau"),
+    "TL": ("TLS", "East Timor"),
+    "TM": ("TKM", "Turkmenistan"),
+    "TN": ("TUN", "Tunisia"),
+    "TO": ("TON", "Tonga"),
+    "TR": ("TUR", "Turkey"),
+    "TT": ("TTO", "Trinidad & Tobago"),
+    "TV": ("TUV", "Tuvalu"),
+    "TW": ("TWN", "Taiwan"),
+    "TZ": ("TZA", "Tanzania"),
+    "UA": ("UKR", "Ukraine"),
+    "UG": ("UGA", "Uganda"),
+    "US": ("USA", "United States"),
+    "UY": ("URY", "Uruguay"),
+    "UZ": ("UZB", "Uzbekistan"),
+    "VA": ("VAT", "Vatican City"),
+    "VC": ("VCT", "St. Vincent & Grenadines"),
+    "VE": ("VEN", "Venezuela"),
+    "VG": ("VGB", "British Virgin Islands"),
+    "VI": ("VIR", "U.S. Virgin Islands"),
+    "VN": ("VNM", "Vietnam"),
+    "VU": ("VUT", "Vanuatu"),
+    "WF": ("WLF", "Wallis & Futuna"),
+    "WS": ("WSM", "Samoa"),
+    "YE": ("YEM", "Yemen"),
+    "YT": ("MYT", "Mayotte"),
+    "ZA": ("ZAF", "South Africa"),
+    "ZM": ("ZMB", "Zambia"),
+    "ZW": ("ZWE", "Zimbabwe"),
+}
+
 
 @dataclass
 class LocaleData:
@@ -1426,6 +1673,36 @@ def country(self) -> str:
         """Generate the country name for this locale."""
         return self._data.address.get("country", "United States")
 
+    def country_name(self) -> str:
+        """Generate the standardized English country name for this locale.
+
+        Uses the `COUNTRY_INFO` mapping to return a clean, standardized English name.
+        Falls back to the address data country name if the code is not found in the mapping.
+        """
+        info = COUNTRY_INFO.get(self.country_code)
+        if info is not None:
+            return info[1]
+        return self._data.address.get("country", "United States")
+
+    def country_code_2(self) -> str:
+        """Generate the ISO 3166-1 alpha-2 (2-letter) country code for this locale."""
+        return self.country_code
+
+    def country_code_3(self) -> str:
+        """Generate the ISO 3166-1 alpha-3 (3-letter) country code for this locale.
+
+        Uses the `COUNTRY_INFO` mapping to look up the 3-letter code from the
+        locale's 2-letter code.
+        """
+        info = COUNTRY_INFO.get(self.country_code)
+        if info is not None:
+            return info[0]
+        # Fallback: try reverse lookup in COUNTRY_CODE_MAP for a 3-letter entry
+        for code, normalized in COUNTRY_CODE_MAP.items():
+            if normalized == self.country_code and len(code) == 3:
+                return code
+        return self.country_code
+
     def postcode(self) -> str:
         """Generate a random postal code (coherent with current location context)."""
         location = self._get_current_location()

pointblank/field.py

@@ -44,6 +44,8 @@
         "city",
         "state",
         "country",
+        "country_code_2",
+        "country_code_3",
         "postcode",
         "latitude",
         "longitude",
@@ -854,7 +856,8 @@ def string_field(
     **Personal:** `"name"` (first + last name), `"name_full"` (full name with possible prefix
     or suffix), `"first_name"`, `"last_name"`, `"email"` (realistic email address),
     `"phone_number"`, `"address"` (full street address), `"city"`, `"state"`, `"country"`,
-    `"postcode"`, `"latitude"`, `"longitude"`
+    `"country_code_2"` (ISO 3166-1 alpha-2 code, e.g., `"US"`), `"country_code_3"` (ISO
+    3166-1 alpha-3 code, e.g., `"USA"`), `"postcode"`, `"latitude"`, `"longitude"`
 
     **Business:** `"company"` (company name), `"job"` (job title), `"catch_phrase"`
 

pointblank/generate/generators.py

@@ -113,6 +113,8 @@ def _generate_from_preset(preset: str, generator: LocaleGenerator) -> str:
         "city": generator.city,
         "state": generator.state,
         "country": generator.country,
+        "country_code_2": generator.country_code_2,
+        "country_code_3": generator.country_code_3,
         "postcode": generator.postcode,
         "latitude": generator.latitude,
         "longitude": generator.longitude,

pointblank_mcp_server/pointblank_server.py

@@ -819,9 +819,9 @@ async def profile_dataframe(
     # Sample data if needed
     if sample_size > 0 and df.shape[0] > sample_size:
         if hasattr(df, "sample"):  # pandas
-            df_sample = df.sample(n=sample_size, random_state=42)
+            df_sample = df.sample(n=sample_size, random_state=23)
         else:  # polars
-            df_sample = df.sample(n=sample_size, seed=42)
+            df_sample = df.sample(n=sample_size, seed=23)
         await ctx.report_progress(20, 100, f"Sampling {sample_size} rows for analysis...")
     else:
         df_sample = df