iotables 0.1.0__py3-none-any.whl

iotables/__init__.py

@@ -0,0 +1,43 @@
+import os as _os
+from importlib.metadata import version as _version, PackageNotFoundError as _PackageNotFoundError
+from iotables.globals import DATA_FOLDER as __DATA_FOLDER
+from iotables.globals import IS_WINDOWS as __IS_WINDOWS
+from iotables.globals import FILES_LOG as __FILES_LOG
+from iotables.oecd import OECD
+from iotables.figaro import Figaro
+from iotables.exiobase import ExioBase
+from iotables.utils import remove_downloaded_files
+
+try:
+    __version__ = _version("iotables")
+except _PackageNotFoundError:  # package not installed (e.g. running from source tree)
+    __version__ = "0.0.0"
+
+
+def get_size_data_folder():
+    """Get size of the folder where downloaded files are stored
+
+    Returns:
+        str: Size of folder
+
+    """
+    from pathlib import Path
+
+    def get_size(folder):
+        def human(size):
+            UNITS = ["B", "KB", "MB", "GB", "TB"]
+            HUMANFMT = "{size} {unit}"
+            HUMANRADIX = 1024.
+            for u in UNITS[:-1]:
+                if size < HUMANRADIX:
+                    return HUMANFMT.format(size=round(size, 2), unit=u)
+                size /= HUMANRADIX
+            return HUMANFMT.format(size=round(size, 2), unit=UNITS[-1])
+
+        root_directory = Path(folder)
+        tot_size = sum(f.stat().st_size for f in root_directory.glob('**/*') if f.is_file())
+        return human(tot_size)
+    return get_size(__DATA_FOLDER)
+
+
+_os.makedirs(__DATA_FOLDER, exist_ok=True)

iotables/base_io.py

@@ -0,0 +1,339 @@
+"""Shared analysis engine: derives the coefficient/inverse matrices and runs shocks."""
+from warnings import warn
+from iotables.matrix import Matrix
+from iotables.utils import assert_is_subset
+import matplotlib.pyplot as plt
+from typing import Union, Iterable, Optional
+import numpy as np
+import pandas as pd
+
+
+class IO:
+    """Class for creating standard IO matrices and IO helper functions, meant as a parent
+    class for classes that load IO data.
+    """
+
+    def __init__(self):
+        necessary_attrs = ['Z',
+                           'X',
+                           'V',
+                           'FD',
+                           'FD_GRAN',
+                           'ADD',
+                           'FD_REGION',
+                           'rs',
+                           'sector_name_mapping',
+                           'regions',
+                           'sectors',
+                           'unit',
+                           'demand_items',
+                           'reference',
+                           'contact']
+        assert_is_subset(necessary_attrs, dir(self))
+
+        # Coefficients matrix, replace 0 with 1 to allow inversion
+        x_filled = self.X.copy()
+        x_filled[x_filled == 0] = 1
+
+        # Leontief demand side
+        self.A = Matrix('Technical coefficients',
+                        self.Z / x_filled.flatten(),
+                        self.Z.rows,
+                        self.Z.columns)
+
+        self.L = Matrix('Leontief inverse',
+                        (np.eye(self.rs) - self.A).I,
+                        self.Z.rows,
+                        self.Z.columns)
+
+        # Ghosh supply side
+        self.B = Matrix('Allocation coefficients',
+                        self.Z / x_filled,
+                        self.Z.rows,
+                        self.Z.columns)
+
+        self.G = Matrix('Ghosh inverse',
+                        (np.eye(self.rs) - self.B).I,
+                        self.Z.rows,
+                        self.Z.columns)
+
+    def _shock(self,
+               model: str,
+               shock: Union[int, float, None] = None,
+               regions: Optional[Iterable] = None,
+               sectors: Optional[Iterable] = None,
+               custom_shock_vector: Optional[Iterable] = None):
+        """Calculates new output using Leontief or Ghosh model
+
+        Args:
+            model: leontief or ghosh
+            shock: Shock in percentage of original final demand or primary inputs
+            regions: List of regions to be shocked
+            sectors: List of sectors to be shocked
+            custom_shock_vector: Vector of length regions * sectors with percentage shocks, overrides all other shock
+                                 parameters if supplied
+
+        Returns:
+            Matrix: Shocked output
+        """
+        if custom_shock_vector is not None:
+            shock_vector = np.array(custom_shock_vector).reshape(self.rs, 1)
+        else:
+            if shock is None or regions is None or sectors is None:
+                raise ValueError("Must supply parameters: 'shock', 'regions', 'sectors'")
+
+            assert_is_subset(regions, self.regions)
+            assert_is_subset(sectors, self.sectors)
+
+            shock_vector = np.array([shock if r in regions and s in sectors
+                                     else 0 for r, s in self.X.rows]).reshape(-1, 1)
+        shock_vector = shock_vector.astype('float64')
+        shock_vector /= 100
+        if model == 'leontief':
+            x_new = (self.L @ (self.FD * shock_vector)) + self.X
+        elif model == 'ghosh':
+            x_new = (self.G.T @ (self.V.T * shock_vector)) + self.X
+        else:
+            raise ValueError('model must be leontief or ghosh')
+
+        return x_new
+
+    def _shock_to_df(self,
+                     x_new: Matrix):
+        """Creates a pandas dataframe with columns: region, sector, x and x_new
+
+        Args:
+            x_new: New output
+
+        Returns:
+            pd.DataFrame
+        """
+
+        return pd.DataFrame({'region': [r for r, s in self.X.rows],
+                             'sector': [s for r, s in self.X.rows],
+                             'x': self.X.flatten(),
+                             'x_new': x_new.flatten()})
+
+    def _shock_and_plot(self,
+                        model: str,
+                        shock: Union[int, float, None] = None,
+                        regions: Optional[Iterable] = None,
+                        sectors: Optional[Iterable] = None,
+                        custom_shock_vector: Optional[Iterable] = None,
+                        plot: bool = False,
+                        plot_by: str = 'region',
+                        plot_regions: Optional[Iterable] = None,
+                        show: bool = True,
+                        ):
+        """Executes a Leontief demand or Ghosh supply shock
+
+        Args:
+            model: leontief or ghosh
+            shock: Shock in percentage of original final demand or primary inputs
+            regions: List of regions to be shocked
+            sectors: List of sectors to be shocked
+            custom_shock_vector: Vector of length regions * sectors with percentage shocks, overrides all other shock
+                                 parameters if supplied
+            plot: Plot if true
+            plot_by: region or sector; if sector, the top 20 is shown
+            plot_regions: List of regions to plot the effect for
+            show: Show the plot if true
+
+        Returns:
+            pd.DataFrame: df with shocked output vector if plot is False, else matplotlib fig, ax
+        """
+        x_new = self._shock(model=model, shock=shock, regions=regions, sectors=sectors,
+                            custom_shock_vector=custom_shock_vector)
+
+        if plot:
+            if plot_regions is None:
+                raise ValueError("Please specify 'plot_regions'")
+            fig, ax = self._plot_shock(x_new=x_new, model=model, by=plot_by, regions=plot_regions)
+            if show:
+                plt.show()
+            return fig, ax
+
+        return self._shock_to_df(x_new)
+
+    def _plot_shock(self,
+                    x_new: Matrix,
+                    model: str,
+                    by: str,
+                    regions: Iterable):
+        """Plot shock from a given new output
+
+        Args:
+            x_new: New output
+            model: ghosh or leontief
+            by: region or sector
+            regions: List of regions to plot the effect for
+
+        Returns:
+            fig, ax
+        """
+        if by not in {'region', 'sector'}:
+            raise ValueError("plot_by must be 'region' or 'sector'")
+        assert_is_subset(regions, self.regions)
+
+        df = self._shock_to_df(x_new)
+        df['diff'] = (df.x_new / df.x - 1).fillna(0)
+        df['col'] = self.V.flatten() if model == 'leontief' else self.FD.flatten()
+        df['newcol'] = df['col'] * (1 + df['diff'])
+
+        df = df[df.region.isin(regions)].groupby(by)[['col', 'newcol']].sum(0)
+        if by == 'sector':
+            df.index = df.index.map(self.sector_name_mapping).str[:25]
+        df['diff'] = (100 * (df.newcol / df.col - 1)).fillna(0)
+        df.sort_values('diff', inplace=True, ascending=False)
+        if by == 'sector':
+            if df['diff'].mean() < 0:
+                df = df.tail(20)
+            else:
+                df = df.head(20)
+
+        fig, ax = plt.subplots()
+        ax.barh(df.index,
+                df['diff']
+                )
+
+        if df['diff'].mean() < 0:
+            ax.invert_xaxis()
+        else:
+            ax.invert_yaxis()
+
+        plt.xlabel(f"% change in {'GVA' if model == 'leontief' else 'final demand'}")
+
+        plt.tight_layout()
+
+        return fig, ax
+
+    def leontief_demand_shock(self,
+                              shock: Union[int, float, None] = None,
+                              regions: Optional[Iterable] = None,
+                              sectors: Optional[Iterable] = None,
+                              custom_shock_vector: Optional[Iterable] = None,
+                              plot: bool = False,
+                              plot_by: str = 'region',
+                              plot_regions: Optional[Iterable] = None,
+                              show: bool = True,
+                              ):
+        """Executes a Leontief demand shock
+
+        Args:
+            shock: Shock in percentage of original demand
+            regions: List of regions to be shocked
+            sectors: List of sectors to be shocked
+            custom_shock_vector: Vector of length regions * sectors with percentage shocks, overrides all other shock
+                                 parameters if supplied
+            plot: Plot if true
+            plot_by: region or sector; if sector, the top 20 is shown
+            plot_regions: List of regions to plot the effect for
+            show: Show the plot if true
+
+        Returns:
+            pd.DataFrame: df with shocked output vector if plot is False, else matplotlib fig, ax
+        """
+        return self._shock_and_plot(model='leontief',
+                                    shock=shock,
+                                    regions=regions,
+                                    sectors=sectors,
+                                    custom_shock_vector=custom_shock_vector,
+                                    plot=plot,
+                                    plot_by=plot_by,
+                                    plot_regions=plot_regions,
+                                    show=show)
+
+    def ghosh_supply_shock(self,
+                           shock: Union[int, float, None] = None,
+                           regions: Optional[Iterable] = None,
+                           sectors: Optional[Iterable] = None,
+                           custom_shock_vector: Optional[Iterable] = None,
+                           plot: bool = False,
+                           plot_by: str = 'region',
+                           plot_regions: Optional[Iterable] = None,
+                           show: bool = True,
+                           ):
+        """Executes a Ghosh supply shock
+
+        Args:
+            shock: Shock in percentage of original primary inputs (value added)
+            regions: List of regions to be shocked
+            sectors: List of sectors to be shocked
+            custom_shock_vector: Vector of length regions * sectors with percentage shocks, overrides all other shock
+                                 parameters if supplied
+            plot: Plot if true
+            plot_by: region or sector; if sector, the top 20 is shown
+            plot_regions: List of regions to plot the effect for
+            show: Show the plot if true
+
+        Returns:
+            pd.DataFrame: df with shocked output vector if plot is False, else matplotlib fig, ax
+        """
+        return self._shock_and_plot(model='ghosh',
+                                    shock=shock,
+                                    regions=regions,
+                                    sectors=sectors,
+                                    custom_shock_vector=custom_shock_vector,
+                                    plot=plot,
+                                    plot_by=plot_by,
+                                    plot_regions=plot_regions,
+                                    show=show)
+
+    def get_imports_exports(self,
+                            import_regions: Iterable,
+                            export_regions: Iterable,
+                            import_sectors: Optional[Iterable] = None,
+                            export_sectors: Optional[Iterable] = None,
+                            use_type: str = 'both'):
+        """Get imports and exports between regions and sectors. Use is broken down by intermediate use and final demand,
+           specify 'use_type' to get either of them or both summed together.
+
+        Args:
+            import_regions: List of importing regions
+            export_regions: List of exporting regions
+            import_sectors: List of importing sectors, all sectors by default, only accounts for intermediate import sectors
+            export_sectors: List of exporting sectors, all sectors by default
+            use_type: 'intermediate', 'final', or 'both'
+
+        Returns:
+            float: Sum of trade flow from exporting region-sectors to importing region-sectors
+        """
+        if use_type not in {'intermediate', 'final', 'both'}:
+            raise ValueError("use_type must be 'intermediate', 'final' or 'both'")
+
+        if import_sectors is not None and use_type in {'final', 'both'}:
+            warn('Note that import_sectors only apply to intermediate use, for final use only import_regions is used')
+
+        if import_sectors is None:
+            import_sectors = self.sectors
+        if export_sectors is None:
+            export_sectors = self.sectors
+
+        if isinstance(import_regions, str):
+            import_regions = [import_regions]
+        if isinstance(export_regions, str):
+            export_regions = [export_regions]
+        if isinstance(import_sectors, str):
+            import_sectors = [import_sectors]
+        if isinstance(export_sectors, str):
+            export_sectors = [export_sectors]
+
+        for subset, superset in zip([import_regions, export_regions, import_sectors, export_sectors],
+                                    [self.regions, self.regions, self.sectors, self.sectors]):
+            assert_is_subset(subset, superset)
+
+        if set(import_regions).intersection(export_regions):
+            warn(f'There is overlap between import_regions and export_regions: '
+                 f'{", ".join(set(import_regions).intersection(export_regions))}')
+
+        intermediate = float(self.Z[np.ix_([r in export_regions and s in export_sectors for r, s in self.Z.rows],
+                                           [r in import_regions and s in import_sectors for r, s in
+                                            self.Z.columns])].sum())
+
+        final = float(
+            self.FD_REGION[np.ix_([r in export_regions and s in export_sectors for r, s in self.FD_REGION.rows],
+                                  [r in import_regions for r in self.FD_REGION.columns])].sum())
+
+        return {'intermediate': intermediate,
+                'final': final,
+                'both': intermediate + final}[use_type]

iotables/config.py

@@ -0,0 +1,195 @@
+"""Single source of truth for the data that exists: download links and shapes per database."""
+
+
+def _chunk_links(base, chunks):
+    """Expand ``{filename: (start_year, end_year)}`` into ``{year: url}``.
+
+    OECD distributes each ICIO edition as a handful of multi-year zip archives
+    (e.g. ``2016-2022_SML.zip`` holds one CSV per year), so every year in a range
+    resolves to the same download URL.
+    """
+    return {year: base + filename
+            for filename, (start, end) in chunks.items()
+            for year in range(start, end + 1)}
+
+
+def _circabc_links(kind_tag, ids):
+    """Build CIRCABC anonymous download URLs for Figaro from ``{year: node_id}``.
+
+    Eurostat distributes the Figaro 2025 edition through CIRCABC; every file has a
+    stable, auth-free download URL of the form ``/sd/a/{node_id}/{filename}``.
+    """
+    return {year: f'https://circabc.europa.eu/sd/a/{nid}/matrix_eu-ic-io_{kind_tag}_25ed_{year}.csv'
+            for year, nid in ids.items()}
+
+
+config = {
+    'oecd':
+        {'2021':
+             {'links': _chunk_links('https://webfs-sti.oecd.org/files/STI-PIE/ICIO/2021/', {
+                  'ICIO2021_1995-1999.zip': (1995, 1999),
+                  'ICIO2021_2000-2004.zip': (2000, 2004),
+                  'ICIO2021_2005-2009.zip': (2005, 2009),
+                  'ICIO2021_2010-2014.zip': (2010, 2014),
+                  'ICIO2021_2015-2018.zip': (2015, 2018),
+              }),
+              'regex_id': r'ICIO2021_[0-9]{4}-[0-9]{4}',
+              'num_regions': 71,
+              'num_sectors': 45
+              },
+         '2022-extended':
+             {'links': _chunk_links('https://webfs-sti.oecd.org/files/STI-PIE/ICIO/2023/', {
+                  '1995-2000_EXT.zip': (1995, 2000),
+                  '2001-2005_EXT.zip': (2001, 2005),
+                  '2006-2010_EXT.zip': (2006, 2010),
+                  '2011-2015_EXT.zip': (2011, 2015),
+                  '2016-2020_EXT.zip': (2016, 2020),
+              }),
+              'regex_id': r'2023/[0-9]{4}-[0-9]{4}_EXT',
+              'num_regions': 81,
+              'num_sectors': 45},
+         '2022-small':
+             {'links': _chunk_links('https://webfs-sti.oecd.org/files/STI-PIE/ICIO/2023/', {
+                  '1995-2000_SML.zip': (1995, 2000),
+                  '2001-2005_SML.zip': (2001, 2005),
+                  '2006-2010_SML.zip': (2006, 2010),
+                  '2011-2015_SML.zip': (2011, 2015),
+                  '2016-2020_SML.zip': (2016, 2020),
+              }),
+              'regex_id': r'2023/[0-9]{4}-[0-9]{4}_SML',
+              'num_regions': 77,
+              'num_sectors': 45},
+        '2025-extended':
+             {'links': _chunk_links('https://webfs-sti.oecd.org/files/STI-PIE/ICIO/2025/', {
+                  '1995-2000_EXT.zip': (1995, 2000),
+                  '2001-2005_EXT.zip': (2001, 2005),
+                  '2006-2010_EXT.zip': (2006, 2010),
+                  '2011-2015_EXT.zip': (2011, 2015),
+                  '2016-2022_EXT.zip': (2016, 2022),
+              }),
+              'regex_id': r'2025/[0-9]{4}-[0-9]{4}_EXT',
+              'num_regions': 85,
+              'num_sectors': 50},
+         '2025-regular':
+             {'links': _chunk_links('https://webfs-sti.oecd.org/files/STI-PIE/ICIO/2025/', {
+                  '1995-2000_SML.zip': (1995, 2000),
+                  '2001-2005_SML.zip': (2001, 2005),
+                  '2006-2010_SML.zip': (2006, 2010),
+                  '2011-2015_SML.zip': (2011, 2015),
+                  '2016-2022_SML.zip': (2016, 2022),
+              }),
+              'regex_id': r'2025/[0-9]{4}-[0-9]{4}_SML',
+              'num_regions': 81,
+              'num_sectors': 50}
+
+         },
+
+    'figaro':
+        {
+         '2025':
+             {'links':
+                  {'product-by-product': _circabc_links('prod-by-prod', {
+                  2010: 'f6d2007e-34c9-420a-a106-8c5ea836c49d',
+                  2011: 'd27f7d7c-4006-4c36-8820-7d9990679323',
+                  2012: '7cea8b73-f166-47ff-9c27-f240794b57af',
+                  2013: 'd56c2ed9-0c4f-40a3-9bbc-1aee49f5bd45',
+                  2014: 'b5aa0e38-3d1b-496b-b180-bbd9bab3e153',
+                  2015: 'd037b772-8bf9-424f-97a4-9274d261e6cf',
+                  2016: '66369d21-8262-47d6-ad8a-dc536a2466d0',
+                  2017: 'd6848c38-9569-4848-bb68-4144fed38c71',
+                  2018: '557b9483-023e-497c-96c5-44ce807bd444',
+                  2019: 'fa0e9127-7a2b-4b74-bbad-1bc9e717197b',
+                  2020: '46f5665e-d2e1-4270-95a1-970da3d70d32',
+                  2021: 'a7e2919d-f084-4c35-8b21-e960a854e2bd',
+                  2022: '31484cee-43fe-45d4-a546-209e5898c6dd',
+                  2023: 'e213892a-afac-4d34-83e9-45e8a324e7e8',
+                  }),
+                   'industry-by-industry': _circabc_links('ind-by-ind', {
+                  2010: 'fc80f855-d144-476e-b4bf-5cfba946819c',
+                  2011: '1bcb2624-04ed-43e1-8588-df6680ed352a',
+                  2012: '399671ad-cbb3-493e-ad5f-83e989f1eecc',
+                  2013: 'a2b4746d-1d11-4a44-ab1c-50ac956f0849',
+                  2014: 'beba57b2-2696-497a-b92f-2a5beca724c7',
+                  2015: '1a194b8c-6ea1-4bec-9c73-0cd599febcc3',
+                  2016: '2cdd74fc-0bce-4ae0-8bf2-34d34546d86d',
+                  2017: '4a11c796-4186-4cce-a02f-12d353fc5e59',
+                  2018: '7a57a374-2200-498c-bb5f-cee24202b0b8',
+                  2019: 'c3467617-8a00-44a0-9b6b-ccad8a2ab58d',
+                  2020: '4df668e1-2a8a-4e84-ae57-00309d8bc760',
+                  2021: '6736dea8-da14-450f-b212-a791baf238c8',
+                  2022: 'b20c339d-984f-413c-a499-54ff76beb90c',
+                  2023: '21557f49-1e94-431c-8523-d972fec020b8',
+                  })},
+              'regex_id': r'matrix_eu-ic-io_[a-z-]+_25ed_[0-9]{4}',
+              'num_regions': 50,
+              'num_sectors': 64},
+        },
+    'exiobase':
+        {'3.81':
+            {'links':
+                {'product-by-product': {
+                    1995: 'https://zenodo.org/record/5589597/files/IOT_1995_pxp.zip?download=1',
+                    1996: 'https://zenodo.org/record/5589597/files/IOT_1996_pxp.zip?download=1',
+                    1997: 'https://zenodo.org/record/5589597/files/IOT_1997_pxp.zip?download=1',
+                    1998: 'https://zenodo.org/record/5589597/files/IOT_1998_pxp.zip?download=1',
+                    1999: 'https://zenodo.org/record/5589597/files/IOT_1999_pxp.zip?download=1',
+                    2000: 'https://zenodo.org/record/5589597/files/IOT_2000_pxp.zip?download=1',
+                    2001: 'https://zenodo.org/record/5589597/files/IOT_2001_pxp.zip?download=1',
+                    2002: 'https://zenodo.org/record/5589597/files/IOT_2002_pxp.zip?download=1',
+                    2003: 'https://zenodo.org/record/5589597/files/IOT_2003_pxp.zip?download=1',
+                    2004: 'https://zenodo.org/record/5589597/files/IOT_2004_pxp.zip?download=1',
+                    2005: 'https://zenodo.org/record/5589597/files/IOT_2005_pxp.zip?download=1',
+                    2006: 'https://zenodo.org/record/5589597/files/IOT_2006_pxp.zip?download=1',
+                    2007: 'https://zenodo.org/record/5589597/files/IOT_2007_pxp.zip?download=1',
+                    2008: 'https://zenodo.org/record/5589597/files/IOT_2008_pxp.zip?download=1',
+                    2009: 'https://zenodo.org/record/5589597/files/IOT_2009_pxp.zip?download=1',
+                    2010: 'https://zenodo.org/record/5589597/files/IOT_2010_pxp.zip?download=1',
+                    2011: 'https://zenodo.org/record/5589597/files/IOT_2011_pxp.zip?download=1',
+                    2012: 'https://zenodo.org/record/5589597/files/IOT_2012_pxp.zip?download=1',
+                    2013: 'https://zenodo.org/record/5589597/files/IOT_2013_pxp.zip?download=1',
+                    2014: 'https://zenodo.org/record/5589597/files/IOT_2014_pxp.zip?download=1',
+                    2015: 'https://zenodo.org/record/5589597/files/IOT_2015_pxp.zip?download=1',
+                    2016: 'https://zenodo.org/record/5589597/files/IOT_2016_pxp.zip?download=1',
+                    2017: 'https://zenodo.org/record/5589597/files/IOT_2017_pxp.zip?download=1',
+                    2018: 'https://zenodo.org/record/5589597/files/IOT_2018_pxp.zip?download=1',
+                    2019: 'https://zenodo.org/record/5589597/files/IOT_2019_pxp.zip?download=1',
+                    2020: 'https://zenodo.org/record/5589597/files/IOT_2020_pxp.zip?download=1',
+                    2021: 'https://zenodo.org/record/5589597/files/IOT_2021_pxp.zip?download=1',
+                    2022: 'https://zenodo.org/record/5589597/files/IOT_2022_pxp.zip?download=1',
+                },
+                    'industry-by-industry': {
+                        1995: 'https://zenodo.org/record/5589597/files/IOT_1995_ixi.zip?download=1',
+                        1996: 'https://zenodo.org/record/5589597/files/IOT_1996_ixi.zip?download=1',
+                        1997: 'https://zenodo.org/record/5589597/files/IOT_1997_ixi.zip?download=1',
+                        1998: 'https://zenodo.org/record/5589597/files/IOT_1998_ixi.zip?download=1',
+                        1999: 'https://zenodo.org/record/5589597/files/IOT_1999_ixi.zip?download=1',
+                        2000: 'https://zenodo.org/record/5589597/files/IOT_2000_ixi.zip?download=1',
+                        2001: 'https://zenodo.org/record/5589597/files/IOT_2001_ixi.zip?download=1',
+                        2002: 'https://zenodo.org/record/5589597/files/IOT_2002_ixi.zip?download=1',
+                        2003: 'https://zenodo.org/record/5589597/files/IOT_2003_ixi.zip?download=1',
+                        2004: 'https://zenodo.org/record/5589597/files/IOT_2004_ixi.zip?download=1',
+                        2005: 'https://zenodo.org/record/5589597/files/IOT_2005_ixi.zip?download=1',
+                        2006: 'https://zenodo.org/record/5589597/files/IOT_2006_ixi.zip?download=1',
+                        2007: 'https://zenodo.org/record/5589597/files/IOT_2007_ixi.zip?download=1',
+                        2008: 'https://zenodo.org/record/5589597/files/IOT_2008_ixi.zip?download=1',
+                        2009: 'https://zenodo.org/record/5589597/files/IOT_2009_ixi.zip?download=1',
+                        2010: 'https://zenodo.org/record/5589597/files/IOT_2010_ixi.zip?download=1',
+                        2011: 'https://zenodo.org/record/5589597/files/IOT_2011_ixi.zip?download=1',
+                        2012: 'https://zenodo.org/record/5589597/files/IOT_2012_ixi.zip?download=1',
+                        2013: 'https://zenodo.org/record/5589597/files/IOT_2013_ixi.zip?download=1',
+                        2014: 'https://zenodo.org/record/5589597/files/IOT_2014_ixi.zip?download=1',
+                        2015: 'https://zenodo.org/record/5589597/files/IOT_2015_ixi.zip?download=1',
+                        2016: 'https://zenodo.org/record/5589597/files/IOT_2016_ixi.zip?download=1',
+                        2017: 'https://zenodo.org/record/5589597/files/IOT_2017_ixi.zip?download=1',
+                        2018: 'https://zenodo.org/record/5589597/files/IOT_2018_ixi.zip?download=1',
+                        2019: 'https://zenodo.org/record/5589597/files/IOT_2019_ixi.zip?download=1',
+                        2020: 'https://zenodo.org/record/5589597/files/IOT_2020_ixi.zip?download=1',
+                        2021: 'https://zenodo.org/record/5589597/files/IOT_2021_ixi.zip?download=1',
+                        2022: 'https://zenodo.org/record/5589597/files/IOT_2022_ixi.zip?download=1',
+                    }
+                },
+                'regex_id': r'[A-Z]{3}_[0-9]{4}_[a-z]{3}',
+                'num_regions': {'industry-by-industry': 49, 'product-by-product': 49},
+                'num_sectors': {'industry-by-industry': 163, 'product-by-product': 200}}
+        }
+}