mlnlib 0.2.35__tar.gz

mlnlib-0.2.35/LICENSE.md

@@ -0,0 +1,40 @@
+### MIT License
+
+Copyright (c) 2024 Eszter Bokányi
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+In any academic work, the package should be cited as follows:
+
+Bokányi, E., Kazmina, Y., & de Jong, R. (2023). POPNET MultiLayerNetwork Python Class (Version v2.0) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.10838865
+
+The BibTeX entry for the citation is:
+```
+@software{bokanyi2023popnet,
+  title = {{{POPNET MultiLayerNetwork Python Class}}},
+  author = {Bok\'anyi, Eszter and Kazmina, Yuliia and family=Jong, given=Rachel, prefix=de, useprefix=true},
+  date = {2023},
+  doi = {10.5281/zenodo.10838865},
+  url = {https://zenodo.org/doi/10.5281/zenodo.10838865},
+  urldate = {2024-11-11},
+  abstract = {Library for working with large multilayer networks by the POPNET group.},
+  organization = {Zenodo},
+  version = {v2.0}
+}
+```
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

mlnlib-0.2.35/PKG-INFO

@@ -0,0 +1,126 @@
+Metadata-Version: 2.1
+Name: mlnlib
+Version: 0.2.35
+Summary: Memory-efficient library for loading and working with large unweighted multilayer networks.
+Author-email: Eszter Bokányi <e.bokanyi@uva.nl>
+Project-URL: Homepage, https://github.com/popnet-io/popnet_mln
+Project-URL: Issues, https://github.com/popnet-io/popnet_mln/issues
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Requires-Python: >=3.8
+Description-Content-Type: text/markdown
+License-File: LICENSE.md
+Requires-Dist: pandas>=1.4.4
+Requires-Dist: polars>=0.20.6
+Requires-Dist: numpy>=1.20.3
+Requires-Dist: python-igraph>=0.9.9
+Requires-Dist: networkx>=3.0
+Requires-Dist: scipy>=1.10.1
+
+# MLN
+
+Library for working with large multilayer networks.
+
+Authors:
+* Eszter Bokányi
+* Rachel de Jong
+* Yuliia Kazmina
+
+Contact:
+`e.bokanyi@uva.nl`
+
+The code is distirbuted under the MIT Licence and should be properly attributed and cited upon reuse - see LICENCE.md for details.
+
+## MultiLayeredNetwork class
+
+The `MultiLayeredNetwork` class from `src/mln.py` contains methods and attributes to work with a large
+multilayer network using different edge types and layers efficiently.
+
+The network has to be unweighted, but can be directed.
+
+The network is either loaded from an already saved sparse matrix and a
+node attribute CSV using the following attributes:
+* adjacency_file:  scipy.sparse.csr matrix saved as an `npz` file
+* node_attribute_file: CSV to be loaded as a pd.DataFrame, rows in
+    order of previous matrix rows should contain a column called
+    "label" that is the primary key for node identification
+
+Or in-memory objects can also be given to the constructor:
+* adjacency_matrix: scipy.sparse.csr matrix
+* node_attribute_dataframe: pd.DataFrame, rows in order of previous
+    matrix rows
+
+Pre-stored files with standard names can be called with the shorthand:
+* from_library: str
+Then the npz and csv.gz files will be read from the
+library_path/from_library folder similarly to the first method.
+
+After loading these two key elements, class attributes and methods work
+the same.
+
+The adjacency matrix self.A is stored in a `scipy.sparse.csr_matrix`
+class, that only saves nonzero elements, and on which scipy csgraph
+algorithms run. People are indexed from 0 to N-1, where N is the total
+number of nodes in this network.
+
+Two dictionaries mapping user id (label) to integer node ids (NID) and
+back are created based on the node attribute file in the
+`self.map_label_to_nid` and `self.map_nid_to_label` attributes. These
+mappings only refer to one instance of the class - if you create a
+subgraph (see later) the coding is going to change, and the mapping
+should be found in the new instance that represents the subgraph.
+
+The adjacency matrix contains integers that encode linktypes if viewed
+as binary numbers. Each possible linktype is assigned an integer of the
+form 2\**i. For example, if both type i and type j edge is present
+between two people, then the corresponding value in self.A would be
+2\**i+2\**j. It means that we can test for a certain edgetype using
+bitwise AND operation very cheaply. E.g. a certain element of self.A is
+7, then 7=1+2+4 which means that edgetypes 0,1, and 2 are present
+between the two people, and 7&2 = 2 in Python (it behaves like a mask
+111 & 010 = 010).
+
+scipy.csr matrices are cheap to slice rowwise, but beware, some
+operation that seem straightforward in numpy might be costly (e.g.
+getting random coordinates after each other or colwise slicing)! If
+something is running too long, consult the scipy reference manual.
+
+Node attributes are stored in `self.node_attributes` which is a
+`pandas.DataFrame`. It is possible to store human-readable or longer
+column names in `self.attribute_colnames`, and code tables for used
+variables in  `self.attribute_code_table`, keys of that variable are the
+colnames in the values of `self.attribute_colnames`.
+
+Example loadings:
+
+```
+# FROM FILES
+# import custom class for the network from mln
+import MultiLayeredNetwork
+
+# read the whole network
+popnet = MultiLayeredNetwork(
+    adjacency_file = 'path_to/adjacency.npz',
+    node_attribute_file = 'path_to/attributes.csv.gz'
+)
+
+# FROM MEMORY
+# import custom class for the network
+from mln import MultiLayeredNetwork
+
+# read the whole network
+popnet = MultiLayeredNetwork(
+    adjacency_matrix = A, # NxN scipy.sparse.csr_matrix
+    node_attribute_file = df # pd.DataFrame containing N rows in the order of the matrix
+)
+
+# FROM LIBRARY
+# import custom class for the network
+from mln import MultiLayeredNetwork
+
+# read the whole network
+popnet = MultiLayeredNetwork(
+    from_library="full"
+)
+```

mlnlib-0.2.35/README.md

@@ -0,0 +1,106 @@
+# MLN
+
+Library for working with large multilayer networks.
+
+Authors:
+* Eszter Bokányi
+* Rachel de Jong
+* Yuliia Kazmina
+
+Contact:
+`e.bokanyi@uva.nl`
+
+The code is distirbuted under the MIT Licence and should be properly attributed and cited upon reuse - see LICENCE.md for details.
+
+## MultiLayeredNetwork class
+
+The `MultiLayeredNetwork` class from `src/mln.py` contains methods and attributes to work with a large
+multilayer network using different edge types and layers efficiently.
+
+The network has to be unweighted, but can be directed.
+
+The network is either loaded from an already saved sparse matrix and a
+node attribute CSV using the following attributes:
+* adjacency_file:  scipy.sparse.csr matrix saved as an `npz` file
+* node_attribute_file: CSV to be loaded as a pd.DataFrame, rows in
+    order of previous matrix rows should contain a column called
+    "label" that is the primary key for node identification
+
+Or in-memory objects can also be given to the constructor:
+* adjacency_matrix: scipy.sparse.csr matrix
+* node_attribute_dataframe: pd.DataFrame, rows in order of previous
+    matrix rows
+
+Pre-stored files with standard names can be called with the shorthand:
+* from_library: str
+Then the npz and csv.gz files will be read from the
+library_path/from_library folder similarly to the first method.
+
+After loading these two key elements, class attributes and methods work
+the same.
+
+The adjacency matrix self.A is stored in a `scipy.sparse.csr_matrix`
+class, that only saves nonzero elements, and on which scipy csgraph
+algorithms run. People are indexed from 0 to N-1, where N is the total
+number of nodes in this network.
+
+Two dictionaries mapping user id (label) to integer node ids (NID) and
+back are created based on the node attribute file in the
+`self.map_label_to_nid` and `self.map_nid_to_label` attributes. These
+mappings only refer to one instance of the class - if you create a
+subgraph (see later) the coding is going to change, and the mapping
+should be found in the new instance that represents the subgraph.
+
+The adjacency matrix contains integers that encode linktypes if viewed
+as binary numbers. Each possible linktype is assigned an integer of the
+form 2\**i. For example, if both type i and type j edge is present
+between two people, then the corresponding value in self.A would be
+2\**i+2\**j. It means that we can test for a certain edgetype using
+bitwise AND operation very cheaply. E.g. a certain element of self.A is
+7, then 7=1+2+4 which means that edgetypes 0,1, and 2 are present
+between the two people, and 7&2 = 2 in Python (it behaves like a mask
+111 & 010 = 010).
+
+scipy.csr matrices are cheap to slice rowwise, but beware, some
+operation that seem straightforward in numpy might be costly (e.g.
+getting random coordinates after each other or colwise slicing)! If
+something is running too long, consult the scipy reference manual.
+
+Node attributes are stored in `self.node_attributes` which is a
+`pandas.DataFrame`. It is possible to store human-readable or longer
+column names in `self.attribute_colnames`, and code tables for used
+variables in  `self.attribute_code_table`, keys of that variable are the
+colnames in the values of `self.attribute_colnames`.
+
+Example loadings:
+
+```
+# FROM FILES
+# import custom class for the network from mln
+import MultiLayeredNetwork
+
+# read the whole network
+popnet = MultiLayeredNetwork(
+    adjacency_file = 'path_to/adjacency.npz',
+    node_attribute_file = 'path_to/attributes.csv.gz'
+)
+
+# FROM MEMORY
+# import custom class for the network
+from mln import MultiLayeredNetwork
+
+# read the whole network
+popnet = MultiLayeredNetwork(
+    adjacency_matrix = A, # NxN scipy.sparse.csr_matrix
+    node_attribute_file = df # pd.DataFrame containing N rows in the order of the matrix
+)
+
+# FROM LIBRARY
+# import custom class for the network
+from mln import MultiLayeredNetwork
+
+# read the whole network
+popnet = MultiLayeredNetwork(
+    from_library="full"
+)
+```

mlnlib-0.2.35/pyproject.toml

@@ -0,0 +1,30 @@
+[build-system]
+requires = ["setuptools>=61.0"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "mlnlib"
+version = "0.2.35"
+authors = [
+  { name="Eszter Bokányi", email="e.bokanyi@uva.nl" },
+]
+description = "Memory-efficient library for loading and working with large unweighted multilayer networks."
+readme = "README.md"
+requires-python = ">=3.8"
+classifiers = [
+    "Programming Language :: Python :: 3",
+    "License :: OSI Approved :: MIT License",
+    "Operating System :: OS Independent",
+]
+dependencies = [
+    "pandas>=1.4.4",
+    "polars>=0.20.6",
+    "numpy>=1.20.3",
+    "python-igraph>=0.9.9",
+    "networkx>=3.0",
+    "scipy>=1.10.1"
+]
+
+[project.urls]
+Homepage = "https://github.com/popnet-io/popnet_mln"
+Issues = "https://github.com/popnet-io/popnet_mln/issues"

mlnlib-0.2.35/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+