lca-modeller 0.0.0__tar.gz

lca_modeller-0.0.0/PKG-INFO

@@ -0,0 +1,69 @@
+Metadata-Version: 2.1
+Name: lca-modeller
+Version: 0.0.0
+Summary: Python library to build parametric LCA models from a configuration file. 
+License: GPL-3.0-only
+Keywords: life cycle assessment,LCA,parametric model,prospective LCA
+Author: Félix POLLET
+Author-email: felix.pollet@isae-supaero.fr
+Requires-Python: >=3.10,<3.12
+Classifier: License :: OSI Approved :: GNU General Public License v3 (GPLv3)
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Requires-Dist: bw2calc (>=1.8.2,<2.0.0)
+Requires-Dist: bw2data (>=3.6.6,<4.0.0)
+Requires-Dist: bw2io (>=0.8.12,<0.9.0)
+Requires-Dist: eight (>=1.0.1,<2.0.0) ; sys_platform == "darwin"
+Requires-Dist: ensure (>=1.0.0,<2.0.0)
+Requires-Dist: intel-openmp (==2023.2.0) ; sys_platform == "darwin"
+Requires-Dist: jupyterlab
+Requires-Dist: lca-algebraic (>=1.2,<2.0)
+Requires-Dist: matplotlib (>=3.7,<4.0)
+Requires-Dist: mkl (==2023.2.2) ; sys_platform == "darwin"
+Requires-Dist: premise (>=2.2.3,<3.0.0)
+Requires-Dist: pyvis (>=0.3.0,<0.4.0)
+Requires-Dist: ruamel.yaml (>=0.15,<0.19)
+Requires-Dist: tbb (==2021.10.0) ; sys_platform == "darwin"
+Description-Content-Type: text/markdown
+
+🚀 LCA-Modeller
+----------------
+*LCA-Modeller* offers a streamlined interface to facilitate the creation of **parametric LCA models** with **prospective capabilities**. It builds on the open-source libraries [*lca-algebraic*](https://lca-algebraic.readthedocs.io/) and [*premise*](https://premise.readthedocs.io/), so having a basic understanding of these tools is recommended.
+
+The core functionality of *LCA-Modeller* revolves around reading a user-provided configuration file that defines the LCA model. From this configuration file, *LCA-Modeller* generates a parametric LCA model with *lca-algebraic*, which can then be evaluated for any parameter values using *lca-algebraic*'s built-in functions.
+<br> 
+If prospective scenarios are provided, *premise* is used to adapt the EcoInvent database to future conditions. The parametric LCA model then interpolates the prospective databases to enable the evaluation for any year specified by the user.
+
+Additional features include the definition of custom impact assessment methods and the ability to modify existing activities in the EcoInvent database by adding or updating flows.
+
+
+📦 Installation
+----------------
+**No pip installation is provided at the moment.** <br>
+To install *LCA-Modeller*, setup a separate conda environment:
+```bash
+conda create -n lca_modeller python==3.10
+conda activate lca_modeller
+```
+And git clone the repository. Then, install *LCA-Modeller* using [poetry](https://python-poetry.org/docs/) by setting the current directory to the root of the repository and running:
+```bash
+poetry install
+```
+
+A tutorial notebook is provided in the `notebooks` directory to help you get started with *LCA-Modeller*.
+
+
+✈️ Applications
+----------------
+*LCA-Modeller* is currently being used in the following projects:
+* [AeroMAPS](https://github.com/AeroMAPS/AeroMAPS) : Multidisciplinary Assessment of Prospective Scenarios for air transport.
+* [FAST-UAV](https://github.com/SizingLab/FAST-UAV): Future Aircraft Sizing Tool - Unmanned Aerial Vehicles
+* [FAST-OAD](https://github.com/fast-aircraft-design/FAST-OAD): Future Aircraft Sizing Tool - Overall Aircraft Design
+
+
+🤝 Questions and contributions
+-------------------------
+* Félix POLLET [felix.pollet@isae-supaero.fr](felix.pollet@isae-supaero.fr)
+
+

lca_modeller-0.0.0/README.md

@@ -0,0 +1,39 @@
+🚀 LCA-Modeller
+----------------
+*LCA-Modeller* offers a streamlined interface to facilitate the creation of **parametric LCA models** with **prospective capabilities**. It builds on the open-source libraries [*lca-algebraic*](https://lca-algebraic.readthedocs.io/) and [*premise*](https://premise.readthedocs.io/), so having a basic understanding of these tools is recommended.
+
+The core functionality of *LCA-Modeller* revolves around reading a user-provided configuration file that defines the LCA model. From this configuration file, *LCA-Modeller* generates a parametric LCA model with *lca-algebraic*, which can then be evaluated for any parameter values using *lca-algebraic*'s built-in functions.
+<br> 
+If prospective scenarios are provided, *premise* is used to adapt the EcoInvent database to future conditions. The parametric LCA model then interpolates the prospective databases to enable the evaluation for any year specified by the user.
+
+Additional features include the definition of custom impact assessment methods and the ability to modify existing activities in the EcoInvent database by adding or updating flows.
+
+
+📦 Installation
+----------------
+**No pip installation is provided at the moment.** <br>
+To install *LCA-Modeller*, setup a separate conda environment:
+```bash
+conda create -n lca_modeller python==3.10
+conda activate lca_modeller
+```
+And git clone the repository. Then, install *LCA-Modeller* using [poetry](https://python-poetry.org/docs/) by setting the current directory to the root of the repository and running:
+```bash
+poetry install
+```
+
+A tutorial notebook is provided in the `notebooks` directory to help you get started with *LCA-Modeller*.
+
+
+✈️ Applications
+----------------
+*LCA-Modeller* is currently being used in the following projects:
+* [AeroMAPS](https://github.com/AeroMAPS/AeroMAPS) : Multidisciplinary Assessment of Prospective Scenarios for air transport.
+* [FAST-UAV](https://github.com/SizingLab/FAST-UAV): Future Aircraft Sizing Tool - Unmanned Aerial Vehicles
+* [FAST-OAD](https://github.com/fast-aircraft-design/FAST-OAD): Future Aircraft Sizing Tool - Overall Aircraft Design
+
+
+🤝 Questions and contributions
+-------------------------
+* Félix POLLET [felix.pollet@isae-supaero.fr](felix.pollet@isae-supaero.fr)
+

lca_modeller-0.0.0/lca_modeller/gui/plots.py

@@ -0,0 +1,88 @@
+"""
+Plots for life cycle assessments interpretation
+"""
+from pyvis.network import Network
+import os
+from IPython.display import display, HTML, IFrame
+from lca_algebraic.activity import Activity
+from lca_modeller.helpers import list_processes
+import matplotlib
+import matplotlib.pyplot as plt
+
+USER_DB = 'Foreground DB'
+default_process_tree_filename = 'process_tree.html'
+
+
+def process_tree(model: Activity,
+                 foreground_only: bool = True,
+                 outfile: str = default_process_tree_filename,
+                 colormap: str = 'Pastel2'):
+    """
+    Plots an interactive tree to visualize the activities and exchanges declared in the LCA module.
+    :param model: The model containing activities and exchanges.
+    :param foreground_only: Boolean flag to include only foreground activities.
+    :param outfile: Output filename for the generated tree.
+    :param colormap: Colormap to use for node coloring.
+    """
+    # Init network
+    net = Network(notebook=True, directed=True, layout=True, cdn_resources='remote', filter_menu=False)
+
+    # Get processes hierarchy
+    df = list_processes(model, foreground_only)
+    df['description'] = df['activity'] + '\n' + df['unit'].fillna('')
+
+    # Colors
+    unique_values = df['database'].unique()
+    value_indices = {value: i for i, value in enumerate(unique_values)}
+    norm = matplotlib.colors.Normalize(vmin=0, vmax=len(unique_values), clip=True)
+    mapper = plt.cm.ScalarMappable(norm=norm, cmap=plt.get_cmap(colormap))
+    colors = df['database'].apply(lambda x: matplotlib.colors.to_hex(mapper.to_rgba(value_indices[x])))
+
+    # Populate network
+    edge_data = zip(df['activity'], df['description'], df['parents'], df['amounts'], df['level'], df['database'], colors)
+
+    edges = {}
+    for activity, description, parents, amounts, level, database, color in edge_data:
+        src = activity if database == USER_DB else f"{activity}\n[{database}]"
+        #desc_array = description.split(',')
+        #desc_db = description.split('(')
+        #desc = ',\n'.join(desc_array[:2]) if len(desc_array) > 1 else desc_array[0]
+        #desc += f"\n({desc_db[-1]}" if len(desc_db) > 1 else ''
+        desc = description
+
+        net.add_node(src, desc, title=src, level=level + 1, shape='box', color=color)
+
+        for parent, amount in zip(parents, amounts):
+            if parent:
+                edge_key = (src, parent)
+                if edge_key in edges:  # Multiple edges between the same nodes
+                    edges[edge_key].append(str(amount))
+                else:  # first time edge links act and its parent: create new node for parent
+                    edges[edge_key] = [str(amount)]
+
+    # Add edges to the network with unique labels
+    for (src, dst), labels in edges.items():
+        if len(labels) == 1:
+            net.add_edge(src, dst, label=labels[0], title=f"Exch. 1: {labels[0]}")
+        else:
+            num_edges = len(labels)
+            roundness_step = 0.1 / num_edges  # Adjust the denominator to control spacing
+            for i, label in enumerate(labels):
+                curve_type = 'curvedCW' if i % 2 == 0 else 'curvedCCW'
+                roundness = roundness_step * (i + 1)  # Adjust roundness to avoid excessive curvature
+                net.add_edge(src, dst, label=label, title=f"Exch. {i + 1}: {label}", smooth={'enabled': True, 'type': curve_type, 'roundness': roundness})
+
+    # Set options
+    net.set_edge_smooth('vertical')
+    net.toggle_physics(False)
+    net.show_buttons(filter_=True)
+    net.options.layout.hierarchical.nodeSpacing = 200
+
+    # Save
+    net.save_graph(outfile)
+
+    # Display in Jupyter Notebook
+    display(IFrame(src=os.path.relpath(outfile), width="100%", height=700))
+
+
+

lca_modeller-0.0.0/lca_modeller/helpers.py

@@ -0,0 +1,206 @@
+import lca_algebraic as agb
+import pandas as pd
+from sympy.parsing.sympy_parser import parse_expr
+from sympy.printing.str import StrPrinter
+from sympy import Float, factor
+import bw2data
+
+USER_DB = 'Foreground DB'
+DEFAULT_PROJECT = 'lca_modeller_default_project'
+
+
+def list_processes(model, foreground_only: bool = True, custom_attribute: str = None) -> pd.DataFrame:
+    """
+    Traverses the tree of sub-activities (sub-processes) until the background database is reached.
+    """
+
+    def _recursive_activities(act, activities, units, locations, parents, amounts, levels, dbs, custom_attributes,
+                              parent: str = "", exc: dict = None, level: int = 0):
+
+        if exc is None:
+            exc = {}
+        name = act.as_dict()['name']
+        unit = act.as_dict()['unit']
+        loc = act.as_dict().get('location', "")
+        if loc not in ['GLO', ''] and f'[{loc}]' not in name:
+            name += f' [{loc}]'
+        amount = _getAmountOrFormula(exc)
+        db = act.as_dict()['database']
+        custom_attr = act.as_dict().get(custom_attribute, "")  # get any additional attribute asked by the user
+
+        # Stop BEFORE reaching the first level of background activities
+        if foreground_only and db != USER_DB:
+            return
+
+        if name in activities:  # Multiple parents for the same activity.
+            # TODO: implement better check than just the name and database, e.g. by adding the act object in the lists (but not in the final df)
+            # Populate lists accordingly and stop recursion since the activity has already been processed
+            idx = activities.index(name)
+            if dbs[idx] == db:
+                parents[idx].append(parent)
+                amounts[idx].append(amount)
+                return
+
+        activities.append(name)
+        units.append(unit)
+        locations.append(loc)
+        parents.append([parent])
+        amounts.append([amount])
+        levels.append(level)
+        dbs.append(db)
+        custom_attributes.append(custom_attr)
+
+        # Stop AFTER reaching the first level of background activities
+        if db != USER_DB:
+            return
+
+        for exc in act.exchanges():
+            if agb.base_utils._isOutputExch(exc):  # skip production exchange to only go down the tree
+                continue
+            _recursive_activities(exc.input, activities, units, locations, parents, amounts, levels, dbs, custom_attributes,
+                                  parent=name,
+                                  exc=exc,
+                                  level=level + 1)
+        return
+
+    # Initialize lists
+    activities = []
+    units = []
+    locations = []
+    parents = []
+    amounts = []
+    levels = []
+    dbs = []
+    custom_attributes = []
+
+    # Recursively populate lists
+    _recursive_activities(model, activities, units, locations, parents, amounts, levels, dbs, custom_attributes)
+    data = {'activity': activities,
+            'unit': units,
+            'location': locations,
+            'level': levels,
+            'database': dbs,
+            'parents': parents,
+            'amounts': amounts,
+            }
+    if custom_attribute:
+        data[custom_attribute] = custom_attributes
+
+    # Create DataFrame
+    df = pd.DataFrame(data, index=activities)
+
+    return df
+
+
+def get_parameter(key: str):
+    param = agb.params._param_registry().__getitem__(key)
+    return param
+
+
+def expandParams(param, value=None):
+    """
+    Modified version of expandParams from classes EnumParam and ParamDef from lca_algebraic library.
+    For enum (switch) parameters, returns a dictionary of single enum values as sympy symbols,
+    with only a single one set to 1.
+    For float parameters, returns a dictionary with either the user-provided value or the default parameter value.
+    """
+
+    # Enum (e.g. switch) parameters
+    if param['type'] == 'enum':
+        values = param['values'] + [None]
+
+        # Bad value ?
+        if value not in values:
+            raise Exception("Invalid value %s for param %s. Should be in %s" %
+                            (value, param['name'], str(param['values'])))
+
+        res = dict()
+        for enum_val in values:
+            var_name = "%s_%s" % (param['name'], enum_val if enum_val is not None else "default")
+            res[var_name] = 1.0 if enum_val == value else 0.0
+        return res
+
+    # Float parameters
+    else:
+        if value is None:
+            value = param['default']
+        return {param['name']: value}
+
+
+def completeParamValues(params, param_registry, setDefaults=True):
+    """
+    Modified version of completeParamValues from lca_algebraic library.
+    Sets default values for missing parameters and expand enum params.
+
+    Returns
+    -------
+        Dict of param_name => float value
+    """
+
+    # Set default variables for missing values
+    if setDefaults:
+        for name, param in param_registry.items():
+            if not name in params:
+                params[name] = param['default']
+                agb.warn(
+                    "Required param '%s' was missing, replacing by default value : %s" % (name, str(param['default'])))
+
+    res = dict()
+    for key, val in params.items():
+        if key in param_registry:
+            param = param_registry[key]
+        else:
+            continue
+            # raise Exception("Parameter not found : %s. Valid parameters : %s" % (key, list(param_registry.keys())))
+
+        if isinstance(val, list):
+            newvals = [expandParams(param, val) for val in val]
+            res.update(agb.params._listOfDictToDictOflist(newvals))
+        else:
+            res.update(expandParams(param, val))
+
+    return res
+
+
+def format_number(num, precision=2):
+    """
+    This function takes a number (either a Python float or a SymPy Float) and a precision,
+    and returns a string that represents the number with the given precision.
+    If the number is too large or too small, it switches to scientific notation.
+
+    :param num: The number to format.
+    :param precision: The number of decimal places to use.
+    :return: A string that represents the number with the given precision.
+    """
+    if isinstance(num, Float):  # SymPy float
+        num = num.evalf()
+    sci_num = "{:.{}e}".format(num, precision)
+    if 'e+00' in sci_num:
+        return "{:.{}f}".format(num, precision)
+    else:
+        return sci_num
+
+
+class CustomStrPrinter(StrPrinter):
+    def _print_Float(self, expr):
+        return '{:.2e}'.format(expr)
+
+
+def _getAmountOrFormula(ex):
+    """ Return either a fixed float value or an expression for the amount of this exchange"""
+    if 'formula' in ex:
+        expr = parse_expr(ex['formula'], evaluate=False)
+        return CustomStrPrinter().doprint(factor(expr))
+    elif 'amount' in ex:
+        return format_number(ex['amount'])
+    return ""
+
+
+def safe_delete_brightway_project(projectname: str) -> None:
+    try:
+        bw2data.projects.delete_project(
+            name = projectname,
+            delete_dir = True
+        )
+    except:
+        pass