Home page

Continuation LF

Continuation often holds between two EDUS when they both elaborate or provide background to the same segment.

LL (linguistic source -> linguistic target)

Rule(s):

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
def LF_Cont_L_L(row):
    if ((row.source_addressee == row.target_addressee 
        and row.distance < 3 
        and (len(row.source_text.split(' ')) < 6 
             and (len(row.target_text.split(' ')) < 3 or row.target_text.split(' ')[0].lower() == 'and')) 
        )
        or ( any( x in row.source_text.lower() for x in ["mmm", "yay", "arrrrrgh", "lulz", "sorry", "well", "anyhoo", "man","yup","ouch", "lol", "oh", "ugh", "no worries","cool", "haha" , "fingers crossed", "actually no","whoa", "wow", "yeah","yep"]) 
            and ((row.target_surface_act == "Question") or (row.target_surface_act == "Assertion")) and (row.distance <= 3)  )
    ):
            l=1
    else:
            l=0
    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
#1 -- get dialogues
finals = []
dialogues = cands.dialogue_num.drop_duplicates()

for d in tqdm(dialogues):
 
    # 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]
            
        #since we know no Result relations are backwards, eliminate immediately   
        backwards = cands[(cands.source_id == s[1]) & (cands.target_id == s[0])]
        if backwards.shape[0] > 0:
            b_index = backwards.index[0]
            finals.append((b_index, -1))
            
        
        #we only look at ling --> ling cases
        
        if row.source_type != 'Segment' or row.target_type != 'Segment':
            
            link = 0  
            
        else: 
        
            link_one = LF_Cont_L_L(row)
            
            if link_one:
                
                link = 1
                
            else:
                
                link = 0
   
                    
        finals.append((r_index, link))

NLNL (non-linguistic source -> non-linguistic target)

Rule(s):

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
def LF_Cont_NL_NL(row):
    if ((willMoveTheRobber(row.source_text) 
        and itsXsTurnToRollTheDice(row.target_text) 
        and row.source_emitter == row.target_emitter 
        and row.distance < 6)
        or 
        (row.distance == 1 and row.source_emitter == row.target_emitter 
         and 
         ((get(row.source_text) and get(row.target_text)) or 
         (has_resources(row.source_text) and has_resources(row.target_text)) or 
         (hasNpoints(row.source_text) and hasNpoints(row.target_text))))
        ):
            l=1
    else:
            l=0
    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
#1 -- get dialogues
finals = []
dialogues = cands.dialogue_num.drop_duplicates()

for d in tqdm(dialogues):
 
    # 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 also 0
            
        backwards = cands[(cands.source_id == s[1]) & (cands.target_id == s[0])]
        if backwards.shape[0] > 0:
            b_index = backwards.index[0]
            finals.append((b_index, -1))
            
        
        #we only look at ling --> ling cases
        
        if row.source_type != 'NonplayerSegment' or row.target_type != 'NonplayerSegment':
            
            link = 0  
            
        else: 
        
            link_one = LF_Cont_NL_NL(row)
            
            if link_one:
                
                link = 1
                
            else:
                
                link = 0
   
                    
        finals.append((r_index, link))