diff utils.py @ 6:e94dc7945639 draft default tip

planemo upload for repository https://github.com/bgruening/galaxytools/tree/recommendation_training/tools/tool_recommendation_model commit 24bab7a797f53fe4bcc668b18ee0326625486164

--- a/utils.py	Fri May 06 09:05:18 2022 +0000
+++ b/utils.py	Sun Oct 16 11:52:10 2022 +0000
@@ -1,11 +1,15 @@
 import json
+import os
 import random
 
 import h5py
 import numpy as np
+import pandas as pd
 import tensorflow as tf
-from numpy.random import choice
-from tensorflow.keras import backend
+
+binary_ce = tf.keras.losses.BinaryCrossentropy()
+binary_acc = tf.keras.metrics.BinaryAccuracy()
+categorical_ce = tf.keras.metrics.CategoricalCrossentropy(from_logits=True)
 
 
 def read_file(file_path):
@@ -17,6 +21,43 @@
     return file_content
 
 
+def write_file(file_path, content):
+    """
+    Write a file
+    """
+    remove_file(file_path)
+    with open(file_path, "w") as json_file:
+        json_file.write(json.dumps(content))
+
+
+def save_h5_data(inp, tar, filename):
+    hf_file = h5py.File(filename, 'w')
+    hf_file.create_dataset("input", data=inp)
+    hf_file.create_dataset("target", data=tar)
+    hf_file.close()
+
+
+def get_low_freq_te_samples(te_data, te_target, tr_freq_dict):
+    lowest_tool_te_ids = list()
+    lowest_t_ids = get_lowest_tools(tr_freq_dict)
+    for i, te_labels in enumerate(te_target):
+        tools_pos = np.where(te_labels > 0)[0]
+        tools_pos = [str(int(item)) for item in tools_pos]
+        intersection = list(set(tools_pos).intersection(set(lowest_t_ids)))
+        if len(intersection) > 0:
+            lowest_tool_te_ids.append(i)
+            lowest_t_ids = [item for item in lowest_t_ids if item not in intersection]
+    return lowest_tool_te_ids
+
+
+def save_processed_workflows(file_path, unique_paths):
+    workflow_paths_unique = ""
+    for path in unique_paths:
+        workflow_paths_unique += path + "\n"
+    with open(file_path, "w") as workflows_file:
+        workflows_file.write(workflow_paths_unique)
+
+
 def format_tool_id(tool_link):
     """
     Extract tool id from tool link
@@ -26,158 +67,166 @@
     return tool_id
 
 
-def set_trained_model(dump_file, model_values):
-    """
-    Create an h5 file with the trained weights and associated dicts
-    """
-    hf_file = h5py.File(dump_file, "w")
-    for key in model_values:
-        value = model_values[key]
-        if key == "model_weights":
-            for idx, item in enumerate(value):
-                w_key = "weight_" + str(idx)
-                if w_key in hf_file:
-                    hf_file.modify(w_key, item)
-                else:
-                    hf_file.create_dataset(w_key, data=item)
-        else:
-            if key in hf_file:
-                hf_file.modify(key, json.dumps(value))
-            else:
-                hf_file.create_dataset(key, data=json.dumps(value))
+def save_model_file(model, r_dict, c_wts, c_tools, s_conn, model_file):
+    model.save_weights(model_file, save_format="h5")
+    hf_file = h5py.File(model_file, 'r+')
+    model_values = {
+        "reverse_dict": r_dict,
+        "class_weights": c_wts,
+        "compatible_tools": c_tools,
+        "standard_connections": s_conn
+    }
+    for k in model_values:
+        hf_file.create_dataset(k, data=json.dumps(model_values[k]))
     hf_file.close()
 
 
-def weighted_loss(class_weights):
-    """
-    Create a weighted loss function. Penalise the misclassification
-    of classes more with the higher usage
-    """
-    weight_values = list(class_weights.values())
-    weight_values.extend(weight_values)
+def remove_file(file_path):
+    if os.path.exists(file_path):
+        os.remove(file_path)
+
 
-    def weighted_binary_crossentropy(y_true, y_pred):
-        # add another dimension to compute dot product
-        expanded_weights = tf.expand_dims(weight_values, axis=-1)
-        bce = backend.binary_crossentropy(y_true, y_pred)
-        return backend.dot(bce, expanded_weights)
-
-    return weighted_binary_crossentropy
+def verify_oversampling_freq(oversampled_tr_data, rev_dict):
+    """
+    Compute the frequency of tool sequences after oversampling
+    """
+    freq_dict = dict()
+    freq_dict_names = dict()
+    for tr_data in oversampled_tr_data:
+        t_pos = np.where(tr_data > 0)[0]
+        last_tool_id = str(int(tr_data[t_pos[-1]]))
+        if last_tool_id not in freq_dict:
+            freq_dict[last_tool_id] = 0
+            freq_dict_names[rev_dict[int(last_tool_id)]] = 0
+        freq_dict[last_tool_id] += 1
+        freq_dict_names[rev_dict[int(last_tool_id)]] += 1
+    s_freq = dict(sorted(freq_dict_names.items(), key=lambda kv: kv[1], reverse=True))
+    return s_freq
 
 
-def balanced_sample_generator(
-    train_data, train_labels, batch_size, l_tool_tr_samples, reverse_dictionary
-):
-    while True:
-        dimension = train_data.shape[1]
-        n_classes = train_labels.shape[1]
-        tool_ids = list(l_tool_tr_samples.keys())
-        random.shuffle(tool_ids)
-        generator_batch_data = np.zeros([batch_size, dimension])
-        generator_batch_labels = np.zeros([batch_size, n_classes])
-        generated_tool_ids = choice(tool_ids, batch_size)
-        for i in range(batch_size):
-            random_toolid = generated_tool_ids[i]
-            sample_indices = l_tool_tr_samples[str(random_toolid)]
-            random_index = random.sample(range(0, len(sample_indices)), 1)[0]
-            random_tr_index = sample_indices[random_index]
-            generator_batch_data[i] = train_data[random_tr_index]
-            generator_batch_labels[i] = train_labels[random_tr_index]
-        yield generator_batch_data, generator_batch_labels
+def collect_sampled_tool_freq(collected_dict, c_freq):
+    for t in c_freq:
+        if t not in collected_dict:
+            collected_dict[t] = int(c_freq[t])
+        else:
+            collected_dict[t] += int(c_freq[t])
+    return collected_dict
+
+
+def save_data_as_dict(f_dict, r_dict, inp, tar, save_path):
+    inp_tar = dict()
+    for index, (i, t) in enumerate(zip(inp, tar)):
+        i_pos = np.where(i > 0)[0]
+        i_seq = ",".join([str(int(item)) for item in i[1:i_pos[-1] + 1]])
+        t_pos = np.where(t > 0)[0]
+        t_seq = ",".join([str(int(item)) for item in t[1:t_pos[-1] + 1]])
+        if i_seq not in inp_tar:
+            inp_tar[i_seq] = list()
+        inp_tar[i_seq].append(t_seq)
+    size = 0
+    for item in inp_tar:
+        size += len(inp_tar[item])
+    print("Size saved file: ", size)
+    write_file(save_path, inp_tar)
+
+
+def read_train_test(datapath):
+    file_obj = h5py.File(datapath, 'r')
+    data_input = np.array(file_obj["input"])
+    data_target = np.array(file_obj["target"])
+    return data_input, data_target
 
 
-def compute_precision(
-    model,
-    x,
-    y,
-    reverse_data_dictionary,
-    usage_scores,
-    actual_classes_pos,
-    topk,
-    standard_conn,
-    last_tool_id,
-    lowest_tool_ids,
-):
-    """
-    Compute absolute and compatible precision
-    """
-    pred_t_name = ""
-    top_precision = 0.0
-    mean_usage = 0.0
-    usage_wt_score = list()
-    pub_precision = 0.0
-    lowest_pub_prec = 0.0
-    lowest_norm_prec = 0.0
-    pub_tools = list()
-    actual_next_tool_names = list()
-    test_sample = np.reshape(x, (1, len(x)))
+def sample_balanced_tr_y(x_seqs, y_labels, ulabels_tr_y_dict, b_size, tr_t_freq, prev_sel_tools):
+    batch_y_tools = list(ulabels_tr_y_dict.keys())
+    random.shuffle(batch_y_tools)
+    label_tools = list()
+    rand_batch_indices = list()
+    sel_tools = list()
+
+    unselected_tools = [t for t in batch_y_tools if t not in prev_sel_tools]
+    rand_selected_tools = unselected_tools[:b_size]
 
-    # predict next tools for a test path
-    prediction = model.predict(test_sample, verbose=0)
-
-    # divide the predicted vector into two halves - one for published and
-    # another for normal workflows
-    nw_dimension = prediction.shape[1]
-    half_len = int(nw_dimension / 2)
+    for l_tool in rand_selected_tools:
+        seq_indices = ulabels_tr_y_dict[l_tool]
+        random.shuffle(seq_indices)
+        rand_s_index = np.random.randint(0, len(seq_indices), 1)[0]
+        rand_sample = seq_indices[rand_s_index]
+        sel_tools.append(l_tool)
+        rand_batch_indices.append(rand_sample)
+        label_tools.append(l_tool)
 
-    # predict tools
-    prediction = np.reshape(prediction, (nw_dimension,))
-    # get predictions of tools from published workflows
-    standard_pred = prediction[:half_len]
-    # get predictions of tools from normal workflows
-    normal_pred = prediction[half_len:]
+    x_batch_train = x_seqs[rand_batch_indices]
+    y_batch_train = y_labels[rand_batch_indices]
 
-    standard_prediction_pos = np.argsort(standard_pred, axis=-1)
-    standard_topk_prediction_pos = standard_prediction_pos[-topk]
+    unrolled_x = tf.convert_to_tensor(x_batch_train, dtype=tf.int64)
+    unrolled_y = tf.convert_to_tensor(y_batch_train, dtype=tf.int64)
+    return unrolled_x, unrolled_y, sel_tools
+
 
-    normal_prediction_pos = np.argsort(normal_pred, axis=-1)
-    normal_topk_prediction_pos = normal_prediction_pos[-topk]
+def sample_balanced_te_y(x_seqs, y_labels, ulabels_tr_y_dict, b_size):
+    batch_y_tools = list(ulabels_tr_y_dict.keys())
+    random.shuffle(batch_y_tools)
+    label_tools = list()
+    rand_batch_indices = list()
+    sel_tools = list()
+    for l_tool in batch_y_tools:
+        seq_indices = ulabels_tr_y_dict[l_tool]
+        random.shuffle(seq_indices)
+        rand_s_index = np.random.randint(0, len(seq_indices), 1)[0]
+        rand_sample = seq_indices[rand_s_index]
+        sel_tools.append(l_tool)
+        if rand_sample not in rand_batch_indices:
+            rand_batch_indices.append(rand_sample)
+            label_tools.append(l_tool)
+        if len(rand_batch_indices) == b_size:
+            break
+    x_batch_train = x_seqs[rand_batch_indices]
+    y_batch_train = y_labels[rand_batch_indices]
+
+    unrolled_x = tf.convert_to_tensor(x_batch_train, dtype=tf.int64)
+    unrolled_y = tf.convert_to_tensor(y_batch_train, dtype=tf.int64)
+    return unrolled_x, unrolled_y, sel_tools
+
 
-    # get true tools names
-    for a_t_pos in actual_classes_pos:
-        if a_t_pos > half_len:
-            t_name = reverse_data_dictionary[int(a_t_pos - half_len)]
-        else:
-            t_name = reverse_data_dictionary[int(a_t_pos)]
-        actual_next_tool_names.append(t_name)
-    last_tool_name = reverse_data_dictionary[x[-1]]
-    # compute scores for published recommendations
-    if standard_topk_prediction_pos in reverse_data_dictionary:
-        pred_t_name = reverse_data_dictionary[int(standard_topk_prediction_pos)]
-        if last_tool_name in standard_conn:
-            pub_tools = standard_conn[last_tool_name]
-            if pred_t_name in pub_tools:
-                pub_precision = 1.0
-                # count precision only when there is actually true published tools
-                if last_tool_id in lowest_tool_ids:
-                    lowest_pub_prec = 1.0
-                else:
-                    lowest_pub_prec = np.nan
-                if standard_topk_prediction_pos in usage_scores:
-                    usage_wt_score.append(
-                        np.log(usage_scores[standard_topk_prediction_pos] + 1.0)
-                    )
-        else:
-            # count precision only when there is actually true published tools
-            # else set to np.nan. Set to 0 only when there is wrong prediction
-            pub_precision = np.nan
-            lowest_pub_prec = np.nan
-    # compute scores for normal recommendations
-    if normal_topk_prediction_pos in reverse_data_dictionary:
-        pred_t_name = reverse_data_dictionary[int(normal_topk_prediction_pos)]
-        if pred_t_name in actual_next_tool_names:
-            if normal_topk_prediction_pos in usage_scores:
-                usage_wt_score.append(
-                    np.log(usage_scores[normal_topk_prediction_pos] + 1.0)
-                )
-            top_precision = 1.0
-            if last_tool_id in lowest_tool_ids:
-                lowest_norm_prec = 1.0
-            else:
-                lowest_norm_prec = np.nan
-    if len(usage_wt_score) > 0:
-        mean_usage = np.mean(usage_wt_score)
-    return mean_usage, top_precision, pub_precision, lowest_pub_prec, lowest_norm_prec
+def get_u_tr_labels(y_tr):
+    labels = list()
+    labels_pos_dict = dict()
+    for i, item in enumerate(y_tr):
+        label_pos = np.where(item > 0)[0]
+        labels.extend(label_pos)
+        for label in label_pos:
+            if label not in labels_pos_dict:
+                labels_pos_dict[label] = list()
+            labels_pos_dict[label].append(i)
+    u_labels = list(set(labels))
+    for item in labels_pos_dict:
+        labels_pos_dict[item] = list(set(labels_pos_dict[item]))
+    return u_labels, labels_pos_dict
+
+
+def compute_loss(y_true, y_pred, class_weights=None):
+    y_true = tf.cast(y_true, dtype=tf.float32)
+    loss = binary_ce(y_true, y_pred)
+    categorical_loss = categorical_ce(y_true, y_pred)
+    if class_weights is None:
+        return tf.reduce_mean(loss), categorical_loss
+    return tf.tensordot(loss, class_weights, axes=1), categorical_loss
+
+
+def compute_acc(y_true, y_pred):
+    return binary_acc(y_true, y_pred)
+
+
+def validate_model(te_x, te_y, te_batch_size, model, f_dict, r_dict, ulabels_te_dict, tr_labels, lowest_t_ids):
+    te_x_batch, y_train_batch, _ = sample_balanced_te_y(te_x, te_y, ulabels_te_dict, te_batch_size)
+    print("Total test data size: ", te_x.shape, te_y.shape)
+    print("Batch test data size: ", te_x_batch.shape, y_train_batch.shape)
+    te_pred_batch, _ = model(te_x_batch, training=False)
+    test_err, _ = compute_loss(y_train_batch, te_pred_batch)
+    print("Test loss:")
+    print(test_err.numpy())
+    print("Test finished")
 
 
 def get_lowest_tools(l_tool_freq, fraction=0.25):
@@ -187,98 +236,7 @@
     return lowest_ids
 
 
-def verify_model(
-    model,
-    x,
-    y,
-    reverse_data_dictionary,
-    usage_scores,
-    standard_conn,
-    lowest_tool_ids,
-    topk_list=[1, 2, 3],
-):
-    """
-    Verify the model on test data
-    """
-    print("Evaluating performance on test data...")
-    print("Test data size: %d" % len(y))
-    size = y.shape[0]
-    precision = np.zeros([len(y), len(topk_list)])
-    usage_weights = np.zeros([len(y), len(topk_list)])
-    epo_pub_prec = np.zeros([len(y), len(topk_list)])
-    epo_lowest_tools_pub_prec = list()
-    epo_lowest_tools_norm_prec = list()
-    lowest_counter = 0
-    # loop over all the test samples and find prediction precision
-    for i in range(size):
-        lowest_pub_topk = list()
-        lowest_norm_topk = list()
-        actual_classes_pos = np.where(y[i] > 0)[0]
-        test_sample = x[i, :]
-        last_tool_id = str(int(test_sample[-1]))
-        for index, abs_topk in enumerate(topk_list):
-            (
-                usg_wt_score,
-                absolute_precision,
-                pub_prec,
-                lowest_p_prec,
-                lowest_n_prec,
-            ) = compute_precision(
-                model,
-                test_sample,
-                y,
-                reverse_data_dictionary,
-                usage_scores,
-                actual_classes_pos,
-                abs_topk,
-                standard_conn,
-                last_tool_id,
-                lowest_tool_ids,
-            )
-            precision[i][index] = absolute_precision
-            usage_weights[i][index] = usg_wt_score
-            epo_pub_prec[i][index] = pub_prec
-            lowest_pub_topk.append(lowest_p_prec)
-            lowest_norm_topk.append(lowest_n_prec)
-        epo_lowest_tools_pub_prec.append(lowest_pub_topk)
-        epo_lowest_tools_norm_prec.append(lowest_norm_topk)
-        if last_tool_id in lowest_tool_ids:
-            lowest_counter += 1
-    mean_precision = np.mean(precision, axis=0)
-    mean_usage = np.mean(usage_weights, axis=0)
-    mean_pub_prec = np.nanmean(epo_pub_prec, axis=0)
-    mean_lowest_pub_prec = np.nanmean(epo_lowest_tools_pub_prec, axis=0)
-    mean_lowest_norm_prec = np.nanmean(epo_lowest_tools_norm_prec, axis=0)
-    return (
-        mean_usage,
-        mean_precision,
-        mean_pub_prec,
-        mean_lowest_pub_prec,
-        mean_lowest_norm_prec,
-        lowest_counter,
-    )
-
-
-def save_model(
-    results,
-    data_dictionary,
-    compatible_next_tools,
-    trained_model_path,
-    class_weights,
-    standard_connections,
-):
-    # save files
-    trained_model = results["model"]
-    best_model_parameters = results["best_parameters"]
-    model_config = trained_model.to_json()
-    model_weights = trained_model.get_weights()
-    model_values = {
-        "data_dictionary": data_dictionary,
-        "model_config": model_config,
-        "best_parameters": best_model_parameters,
-        "model_weights": model_weights,
-        "compatible_tools": compatible_next_tools,
-        "class_weights": class_weights,
-        "standard_connections": standard_connections,
-    }
-    set_trained_model(trained_model_path, model_values)
+def remove_pipe(file_path):
+    dataframe = pd.read_csv(file_path, sep="|", header=None)
+    dataframe = dataframe[1:len(dataframe.index) - 1]
+    return dataframe[1:]