Dream emotion dist

annotators/bot_emotion_classifier/Dockerfile

@@ -0,0 +1,17 @@
+FROM python:3.9.16
+WORKDIR /src
+
+COPY annotators/bot_emotion_classifier/requirements.txt ./requirements.txt
+RUN pip install -r requirements.txt
+
+ARG SERVICE_PORT
+ENV SERVICE_PORT=$SERVICE_PORT
+
+COPY annotators/bot_emotion_classifier/ ./
+COPY common /src/common
+RUN python -c 'import stanza; stanza.download("en")'
+
+# wait for a server answer ( INTERVAL + TIMEOUT ) * RETRIES seconds after that change status to unhealthy
+HEALTHCHECK --interval=5s --timeout=5s --retries=3 CMD curl --fail 127.0.0.1:${SERVICE_PORT}/healthcheck || exit 1
+
+CMD gunicorn --workers=1 server:app -b 0.0.0.0:${SERVICE_PORT} --timeout=600

annotators/bot_emotion_classifier/dscript_scheme_classifier.py

@@ -0,0 +1,139 @@
+import stanza
+
+stanza.download("en")
+nlp = stanza.Pipeline("en")
+
+
+pos_type_expl = {
+    "i - someone": 1,
+    "you - someone": 2,
+    "i - you": 3,
+    "you - you": 3,
+    "i - i": 4,
+    "you - i": 5,
+    "we - someone": 6,
+    "someone - someone": 6,
+    "we - we": 7,
+}
+
+neg_type_expl = {
+    "i - you": 1,
+    "i - someone": 2,
+    "we - someone": 3,
+    "you - someone": 3,
+    "you - i": 4,
+    "i - i": 5,
+    "we - we": 5,
+}
+
+first_prons = ["i", "me", "myself"]
+second_prons = ["you"]
+inclusive_prons = ["we"]
+
+
+def find_root(sent):
+    for token in sent:
+        if token["deprel"] == "root":
+            return token
+
+
+def find_clause_head(root, sent):
+    clauses = ["ccomp", "xcomp", "acl", "acl:relcl", "advcl"]
+    all_clause_heads = []
+    head_id = root["id"]
+    for token in sent:
+        if token["head"] == head_id and token["deprel"] in clauses:
+            all_clause_heads.append(token)
+    return all_clause_heads
+
+
+def find_arguments(head, sent):
+    objects = ["obl", "obj", "iobj"]
+    head_id = head["id"]
+    subj = ""
+    obj = ""
+    for token in sent:
+        if token["head"] == head_id and "subj" in token["deprel"]:
+            subj = token
+        elif token["head"] == head_id and token["deprel"] in objects:
+            obj = token
+    return subj, obj
+
+
+def reverse_if_not_verb(root, subj, obj, has_clauses):
+    not_verbs = ["NOUN", "ADJ", "ADV"]
+    if has_clauses:
+        return subj, obj
+    if root["upos"] in not_verbs:
+        obj = subj
+        subj = "I"
+    return subj, obj
+
+
+def find_final_arguments(sent):
+    root = find_root(sent)
+    subj, obj = find_arguments(root, sent)
+    next_clause_heads = find_clause_head(root, sent)
+    has_clauses = False
+    if next_clause_heads:
+        has_clauses = True
+    queue = next_clause_heads
+
+    if subj and not obj:
+        dep_subj, dep_obj = "", ""
+        while not dep_subj and not dep_obj and queue:
+            root = queue[0]
+            queue = queue[1:]
+            dep_subj, dep_obj = find_arguments(root, sent)
+            next_clause_heads = find_clause_head(root, sent)
+            queue.extend(next_clause_heads)
+        if dep_subj:
+            obj = dep_subj
+        else:
+            obj = {"text": "someone"}
+        return reverse_if_not_verb(root, subj["text"], obj["text"], has_clauses)
+
+    while not subj and not obj and queue:
+        root = queue[0]
+        queue = queue[1:]
+        subj, obj = find_arguments(root, sent)
+        next_clause_heads = find_clause_head(root, sent)
+        queue.extend(next_clause_heads)
+
+    if obj and not subj:
+        if "Mood=Imp" in root["feats"]:
+            subj = {"text": "you"}
+        else:
+            subj = {"text": "someone"}
+        return reverse_if_not_verb(root, subj["text"], obj["text"], has_clauses)
+    elif not subj and not obj:
+        subj = {"text": "someone"}
+        obj = {"text": "someone"}
+        return reverse_if_not_verb(root, subj["text"], obj["text"], has_clauses)
+    elif subj and obj:
+        return subj["text"], obj["text"]
+    else:
+        obj = {"text": "someone"}
+        return subj["text"], obj["text"]
+
+
+def get_dsript_type(orig_sent, type_expl):
+    doc = nlp(orig_sent)
+    sent = doc.sentences[0].to_dict()
+    subj, obj = find_final_arguments(sent)
+    subj = subj.lower()
+    obj = obj.lower()
+    if subj not in first_prons and subj not in second_prons and subj not in inclusive_prons:
+        subj = "someone"
+    if obj not in first_prons and obj not in second_prons and obj not in inclusive_prons:
+        obj = "someone"
+    if subj in first_prons:
+        subj = "i"
+    if obj in first_prons:
+        obj = "i"
+    line = subj + " - " + obj
+    if line not in type_expl:
+        type_num = 3
+    else:
+        type_num = type_expl[line]
+    return type_num

annotators/bot_emotion_classifier/info_files/neg_reactions.txt

@@ -0,0 +1,49 @@
+anger	shame
+resentment	distress
+disappointment	shame
+disgust	distress
+shame	shame
+distress	shame
+fear	shame
+sadness	shame
+surprise	shame
+
+anger	pity
+resentment	pity
+disappointment	pity
+disgust	neutral
+shame	pity
+distress	pity
+fear	fear
+sadness	pity
+surprise	pity
+
+anger	anger
+resentment	resentment
+disappointment	disappointment
+disgust	disgust
+shame	shame
+distress	distress
+fear	fear
+sadness	sadness
+surprise	distress
+
+anger	resentment
+resentment	resentment
+disappointment	distress
+disgust	resentment
+shame	resentment
+distress	distress
+fear	fear
+sadness	resentment
+surprise	distress
+
+anger	pity
+resentment	pity
+disappointment	pity
+disgust	pity
+shame	pity
+distress	pity
+fear	pity
+sadness	pity
+surprise	pity

annotators/bot_emotion_classifier/info_files/pos_reactions.txt

@@ -0,0 +1,69 @@
+admiration	admiration
+joy	joy
+liking	liking
+love	admiration
+hope	hope
+gratitude	admiration
+pride	admiration
+relief	relief
+surprise	joy
+
+admiration	admiration
+joy	joy
+liking	liking
+love	love
+hope	hope
+gratitude	gratitude
+pride	pride
+relief	relief
+surprise	joy
+
+admiration	gratitude
+joy	joy
+liking	gratitude
+love	love
+hope	gratitude
+gratitude	joy
+pride	gratitude
+relief	joy
+surprise	joy
+
+admiration	admiration
+joy	joy
+liking	admiration
+love	admiration
+hope	hope
+gratitude	joy
+pride	admiration
+relief	joy
+surprise	joy
+
+admiration	admiration
+joy	joy
+liking	liking
+love	love
+hope	hope
+gratitude	gratitude
+pride	pride
+relief	relief
+surprise	joy
+
+admiration	admiration
+joy	joy
+liking	liking
+love	love
+hope	hope
+gratitude	gratitude
+pride	pride
+relief	relief
+surprise	joy
+
+admiration	admiration
+joy	joy
+liking	admiration
+love	admiration
+hope	hope
+gratitude	gratitude
+pride	pride
+relief	relief
+surprise	joy

annotators/bot_emotion_classifier/requirements.txt

@@ -0,0 +1,14 @@
+attrdict==2.0.1
+flask==1.1.1
+gunicorn==19.9.0
+requests==2.22.0
+numpy==1.17.2
+sentry-sdk[flask]==0.14.1
+stanza==1.2.2
+six==1.16.0
+itsdangerous==2.0.1
+uvicorn==0.13.0
+healthcheck==1.3.3
+requests==2.22.0
+jinja2<=3.0.3
+Werkzeug==2.0.3