

In [2]:

    
import pymongo
from IPython.display import display









    




ImportErrorTraceback (most recent call last)
<ipython-input-2-b4b8a0a73d36> in <module>()
----> 1 import pymongo
      2 from IPython.display import display

ImportError: No module named pymongo



In [4]:

    
client = pymongo.MongoClient("localhost", 27017)









    




NameErrorTraceback (most recent call last)
<ipython-input-4-953460a3945d> in <module>()
----> 1 client = pymongo.MongoClient("localhost", 27017)

NameError: name 'pymongo' is not defined



In [4]:

    
db = client.joint_ner_and_md



In [ ]:

    
print("section1-all-20171013-01")
runs = db.runs.find({"config.experiment_name": "section1-all-20171013-01"})
configs = []
for run_idx, run in enumerate(runs):
    
    dict_to_report = dict(run["config"])
    initial_keys = dict_to_report.keys()
    
    print initial_keys
    
    result_designation_labels = ["MORPH", "NER", "YURET"]
    
    for result_designation_label in result_designation_labels:
        
        print "result_designation_label: ", result_designation_label
    
        if result_designation_label == "YURET":
            best_performances = run["info"][result_designation_label + "_test_f_score"]
        else:
            best_performances = run["info"][result_designation_label + "_dev_f_score"]
        print best_performances
        best_dev_result_for_this_run = 0
        best_test_result_for_this_run = 0
        epoch_id_of_the_best_dev_result = -1
        # display(run["config"])
        for epoch in sorted([int(k) for k in best_performances.keys()]):
            # if result_designation_label != "NER":
            #     corrected_epoch = epoch + 1
            epoch_max = max(best_performances[str(epoch)])
            if epoch_max > best_dev_result_for_this_run:
                epoch_id_of_the_best_dev_result = epoch
                best_dev_result_for_this_run = epoch_max
                best_test_result_for_this_run = \
                    max(run["info"][result_designation_label + "_test_f_score"][str(epoch)])
                
            # print "run_idx: %d, epoch: %d, epoch_best_performance: %.2lf, best_for_this_run: %.2lf" % (run_idx, epoch, epoch_max, best_for_this_run)
    
        dict_to_report[result_designation_label + "_best_dev"] = best_dev_result_for_this_run
        dict_to_report[result_designation_label + "_best_test"] = best_test_result_for_this_run
        
        for x in result_designation_labels:
            # if x != result_designation_label:
            print "x: ", x
            print "epoch_id_of_the_best_dev_result: ", epoch_id_of_the_best_dev_result
            dict_to_report[result_designation_label + "_to_" + x + "_test"] = \
                max(run["info"][x + "_test_f_score"][str(epoch_id_of_the_best_dev_result)]) \
                    if str(epoch_id_of_the_best_dev_result) in run["info"][x + "_test_f_score"].keys() else -1
            print dict_to_report[result_designation_label + "_to_" + x + "_test"]
        
    configs.append({key: dict_to_report[key] for key in ["host", 
                                                         "integration_mode", 
                                                         "train_with_yuret", 
                                                         "use_golden_morpho_analysis_in_word_representation"] + 
                    [x for x in dict_to_report.keys() if x not in initial_keys]})

import pandas
df = pandas.DataFrame.from_dict(configs)
print configs
cols = df.columns.tolist()

# display(df[["host"] + 
#                     [x for x in dict_to_report.keys() if x not in initial_keys]])

display(df)









    Out[ ]:






  
    
      
      char_dim
      morpho_tag_dim
      word_dim
      morpho_tag_type
      host
      best
    
  
  
    
      0
      0
      10
      10
      with_root
      localhost
      59.61