Home page

Comment LF

Comment holds if the target DU provides an opinion or evaluation of the content associated with the first DU.

LL (linguistic source -> linguistic target)

Rule(s):

1
2
3
4
5
6
7
def LF_Com_L_L_case1(row):
    #case 1
    l=0
    if row.distance ==1:
        if (is_comment(row.target_text) and len(row.target_text.split(' ')) <= 6) or (contains_emoji(row.target_text)):
            l=1
    return l

Application of rules to candidates:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
#1 -- get dialogues
finals = []
dialogues = cands.dialogue_num.drop_duplicates()
#dialogues = [97]

for d in tqdm(dialogues):
     
    #assume each target cannot be used more than once 
    target_memo = []
    # 2 -- get segment list
    seg_list = get_seg_list(cands[cands.dialogue_num == d])

    # 3 -- create seg pairs list
    seg_pairs = []
    for i, s in enumerate(seg_list):
        for n in [j for j in reversed(range((i+1)-20, i+1)) if j>=0]:
            try:
                seg_pairs.append((seg_list[n], seg_list[i+1]))   
            except IndexError:
                pass
            
    # 4 -- for each pair, pull row and append rules

    for s in seg_pairs:
        row = cands[(cands.source_id == s[0]) & (cands.target_id == s[1])]
        if row.empty:
            #check for backwards links
            backwards = cands[(cands.source_id == s[1]) & (cands.target_id == s[0])]
            if backwards.shape[0] > 0:
                b_index = backwards.index[0]
                b_row = backwards.iloc[0]
                if (b_row.source_turn_id == b_row.target_turn_id) and LF_Com_L_L_case1(b_row):
                    if b_row.target_id in target_memo:
                        finals.append((b_index, 0))
                    else:
                        finals.append((b_index, 1))
                        target_memo.append(b_row.target_id)
                else:
                    finals.append((b_index, 0))     
            continue
        else:
            r_index = row.index[0]
            row = row.iloc[0]
            backwards = cands[(cands.source_id == s[1]) & (cands.target_id == s[0])]
        
            #we only look at ling --> ling cases
        
            if row.source_type != 'Segment' or row.target_type != 'Segment':
                finals.append((r_index, 0))
                if backwards.shape[0] > 0:
                    b_index = backwards.index[0]
                    finals.append((b_index, -1))
                    
            elif LF_Com_L_L_case1(row):
                if row.target_id in target_memo:
                    finals.append((r_index, 0))
                else:
                    finals.append((r_index, 1))
                    target_memo.append(row.target_id)
                #then no backwards link
                if backwards.shape[0] > 0:
                    b_index = backwards.index[0]
                    finals.append((b_index, -1))
                
            elif backwards.shape[0] > 0:
                b_row = backwards.iloc[0]
                b_index = backwards.index[0]
                if (b_row.source_turn_id == b_row.target_turn_id) and LF_Com_L_L_case1(b_row):
                    if row.target_id in target_memo:
                        finals.append((b_index, 0))
                        finals.append((r_index, 0))
                    else:
                        finals.append((b_index, 1))
                        finals.append((r_index, -1))
                        target_memo.append(row.target_id)
                else:
                    finals.append((b_index, 0))
                    finals.append((r_index, 0))
            
            else:
                finals.append((r_index, 0))

NLL (non-linguistic source -> linguistic target)

Rule(s):

1
2
3
4
5
6
7
8
def LF_Com_NL_L_case1(row):
    #case 1
    l=0
    if row.distance ==1:
        if ((is_comment(row.target_text) and len(row.target_text.split(' ')) <= 6) \
        or (contains_emoji(row.target_text))) and row.target_emitter in row.source_text.split(' '):
            l=1
    return l

Application of rules to candidates:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
#1 -- get dialogues
finals = []
dialogues = cands.dialogue_num.drop_duplicates()

for d in tqdm(dialogues):
    
    target_memo = []
    # 2 -- get segment list
    seg_list = get_seg_list(cands[cands.dialogue_num == d])

    # 3 -- create seg pairs list
    seg_pairs = []
    for i, s in enumerate(seg_list):
        for n in [j for j in reversed(range((i+1)-20, i+1)) if j>=0]:
            try:
                seg_pairs.append((seg_list[n], seg_list[i+1]))   
            except IndexError:
                pass
            
    # 4 -- for each pair, pull row and append rules
    for s in seg_pairs:
        row = cands[(cands.source_id == s[0]) & (cands.target_id == s[1])]
        if row.empty:
            continue
        else:
            r_index = row.index[0]
            row = row.iloc[0]
        
        # make sure that if any backwards links exist they are -1, unless both are "Segments" in 
        # which case we don't know
          
        backwards = cands[(cands.source_id == s[1]) & (cands.target_id == s[0])]
        if backwards.shape[0] > 0:
            b_index = backwards.index[0]
            if row.source_type == 'Segment' and row.target_type == 'Segment':
                finals.append((b_index, 0))
            else:
                finals.append((b_index, -1))
            
        #we only look at ling --> ling cases
        
        if row.source_type != 'NonplayerSegment' or row.target_type != 'Segment':
            
            link = 0  
            
        else: 
            
            #try each of three cases, keeping track of the candidates which have been used in a sequence relation already
        
            if LF_Com_NL_L_case1(row):
                if row.target_id in target_memo:
                    link = 0
                else:    
                    link = 1
                    target_memo.append(row.target_id)
            else:
                link = 0
                    
        finals.append((r_index, link))