Merge pull request #2 from J41R0/dev

Update to version 1.0.0

Author: J41R0

README.md

@@ -1,13 +1,34 @@
-# py-fcm
-Fuzzy cognitive maps python library
+# PyFCM
+Fuzzy cognitive maps python library. Also, supports the topology generation from data to solve classification problems.
+The details associated to the generation process are described in [this paper](https://link.springer.com/chapter/10.1007/978-3-030-89691-1_25). 
+### Installation
 
-###Example usage
+#### From source:
 
+1. Clone repository:
+    ```
+    $ git clone https://github.com/J41R0/PyFCM.git 
+    $ cd PyFCM
+    ```
+2. Install setup tools and package:
+    ```
+    $ pip install setuptools
+    $ python setup.py install
+    ```
+#### From PyPi:
+1. Install package using pip:
+    ```
+    $ pip install py-fcm
+    ```
+   
+### Example usage
+
+#### Inference:
 ```
 from py_fcm import from_json
 
 fcm_json = """{
-         "max_iter": 500,
+            "max_iter": 500,
             "decision_function": "LAST",
             "activation_function": "sigmoid",
             "memory_influence": False,
@@ -54,6 +75,29 @@ fcm_json = """{
         """
 my_fcm = from_json(fcm_json)
 my_fcm.run_inference()
-result = my_fcm.get_result_by_type(node_type='any')
+result = my_fcm.get_final_state(concept_type='any')
 print(result)
+```
+
+#### Generation:
+```
+import pandas
+from py_fcm import FcmEstimator
+
+data_dict = {
+   'F1': ['x', 'x', 'y', 'y'],
+   'F2': [9.8, 7.3, 1.1, 3.6],
+   'class': ['a', 'a', 'r', 'r']
+}
+    
+ train = pandas.DataFrame(data_dict)
+ x_train = train.loc[:, train.columns != 'class']
+ y_train = train.loc[:, 'class']
+
+ estimator = FcmEstimator()
+ estimator.fit(x_train, y_train)
+ print(estimator.predict(x_train))
+ print("Accuracy: ",estimator.score(x_train, y_train))
+ print(estimator.get_fcm().to_json())
+
 ```

py_fcm/__init__.py

@@ -1,2 +1,46 @@
-from py_fcm.fcm import FuzzyCognitiveMap, from_json, join_maps
-from py_fcm.__const import TYPE_SIMPLE, TYPE_DECISION, TYPE_FUZZY
+from py_fcm.utils.functions import Relation
+from py_fcm.fcm_estimator import FcmEstimator
+from py_fcm.loader import from_json, join_maps, FuzzyCognitiveMap
+from py_fcm.utils.__const import TYPE_SIMPLE, TYPE_DECISION, TYPE_FUZZY
+
+
+def load_csv_dataset(dataset_dir, ds_name, factorize=False):
+    import pandas
+    from collections import OrderedDict
+    data_dict = OrderedDict()
+    with open(dataset_dir + "/" + ds_name) as file:
+        content = []
+        for line in file:
+            content.append(line.strip())
+    # first line describe the atributes names
+    attributes = str(content[0]).split(',')
+    for current_att in attributes:
+        data_dict[current_att] = []
+
+    # print(data_dict)
+    for line in range(1, len(content)):
+        data_line = str(content[line]).split(',')
+        # avoid rows with different attributes length
+        if len(data_line) == len(attributes):
+            # the missing data must be identified
+            for data in range(0, len(data_line)):
+                # reusing for value type inference
+                current_att = data_line[data]
+                data_dict[attributes[data]].append(current_att)
+        else:
+            # Handle errors in dataset matrix
+            print("Errors in line: ", line, len(data_line), len(attributes))
+    # adding data set
+    print("\n===> Dataset for test: ", ds_name)
+    try:
+        dataset_frame = pandas.DataFrame(data_dict).drop_duplicates()
+    except Exception as err:
+        for key, value in data_dict.items():
+            print(key, value)
+            pass
+        raise Exception(ds_name + " " + str(err))
+    # Transform the columns disc values in 0..N values
+    if factorize:
+        # for current_col in dataset_frame.
+        dataset_frame = dataset_frame.apply(lambda x: pandas.factorize(x)[0])
+    return dataset_frame