Spaces:
Running
Running
Update modules/morphosyntax/morphosyntax_interface.py
Browse files
modules/morphosyntax/morphosyntax_interface.py
CHANGED
|
@@ -42,216 +42,184 @@ import base64
|
|
| 42 |
import re
|
| 43 |
|
| 44 |
############################################################################
|
|
|
|
| 45 |
def display_morphosyntax_interface(lang_code, nlp_models, morpho_t):
|
| 46 |
try:
|
| 47 |
-
# CSS
|
| 48 |
st.markdown("""
|
| 49 |
<style>
|
| 50 |
.stTextArea textarea {
|
| 51 |
font-size: 1rem;
|
| 52 |
line-height: 1.5;
|
| 53 |
-
padding: 0.5rem;
|
| 54 |
-
border-radius: 0.375rem;
|
| 55 |
-
border: 1px solid #e2e8f0;
|
| 56 |
-
background-color: white;
|
| 57 |
min-height: 100px !important;
|
| 58 |
height: 100px !important;
|
| 59 |
}
|
| 60 |
.arc-diagram-container {
|
| 61 |
width: 100%;
|
| 62 |
-
overflow-x: auto;
|
| 63 |
-
background-color: white;
|
| 64 |
padding: 0.5rem;
|
| 65 |
-
border-radius: 0.375rem;
|
| 66 |
-
box-shadow: 0 2px 4px rgba(0,0,0,0.1);
|
| 67 |
margin: 0.5rem 0;
|
| 68 |
}
|
| 69 |
-
.stTab {
|
| 70 |
-
position: sticky;
|
| 71 |
-
top: 0;
|
| 72 |
-
z-index: 999;
|
| 73 |
-
}
|
| 74 |
-
[data-testid="stVerticalBlock"] {
|
| 75 |
-
gap: 0 !important;
|
| 76 |
-
}
|
| 77 |
</style>
|
| 78 |
""", unsafe_allow_html=True)
|
| 79 |
|
| 80 |
-
#
|
| 81 |
-
if '
|
| 82 |
-
st.session_state.
|
| 83 |
-
|
| 84 |
-
|
| 85 |
-
|
| 86 |
-
|
| 87 |
-
|
| 88 |
-
|
| 89 |
-
'original_analysis': None,
|
| 90 |
-
'current_text': '',
|
| 91 |
-
'current_analysis': None,
|
| 92 |
-
'analysis_count': 0,
|
| 93 |
-
'show_original': False,
|
| 94 |
-
'show_iteration': False
|
| 95 |
-
}
|
| 96 |
-
|
| 97 |
-
for key, default_value in default_state.items():
|
| 98 |
-
if key not in st.session_state.morphosyntax_state:
|
| 99 |
-
st.session_state.morphosyntax_state[key] = default_value
|
| 100 |
-
|
| 101 |
-
def update_analysis():
|
| 102 |
-
st.session_state.current_tab = "morphosyntax" # Mantener tab activo
|
| 103 |
|
| 104 |
-
#
|
| 105 |
-
st.
|
| 106 |
-
|
| 107 |
-
|
| 108 |
-
|
| 109 |
-
|
| 110 |
-
height=100
|
| 111 |
-
on_change=update_analysis
|
| 112 |
)
|
| 113 |
|
|
|
|
| 114 |
col1, col2, col3 = st.columns([2,1,2])
|
| 115 |
with col1:
|
| 116 |
-
|
| 117 |
"Analizar Texto Original",
|
| 118 |
type="primary",
|
| 119 |
-
use_container_width=True
|
| 120 |
-
disabled=not bool(st.session_state.get("original_text_input", "").strip())
|
| 121 |
)
|
| 122 |
|
| 123 |
-
# Procesar
|
| 124 |
-
if
|
| 125 |
try:
|
| 126 |
-
|
| 127 |
-
|
| 128 |
-
|
| 129 |
-
|
| 130 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 131 |
)
|
| 132 |
-
|
| 133 |
-
st.session_state.morphosyntax_state.update({
|
| 134 |
-
'original_text': st.session_state.original_text_input,
|
| 135 |
-
'original_analysis': {'doc': doc, 'analysis': analysis},
|
| 136 |
-
'current_text': st.session_state.original_text_input,
|
| 137 |
-
'show_original': True,
|
| 138 |
-
'analysis_count': st.session_state.morphosyntax_state['analysis_count'] + 1
|
| 139 |
-
})
|
| 140 |
|
| 141 |
-
|
| 142 |
-
|
| 143 |
-
|
| 144 |
-
|
| 145 |
-
|
| 146 |
-
|
|
|
|
|
|
|
| 147 |
|
| 148 |
-
|
| 149 |
-
|
| 150 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 151 |
|
| 152 |
-
|
| 153 |
-
|
| 154 |
-
|
| 155 |
-
|
| 156 |
-
|
| 157 |
-
morpho_t
|
| 158 |
-
)
|
| 159 |
|
| 160 |
-
|
| 161 |
-
|
| 162 |
-
|
| 163 |
-
|
| 164 |
-
|
| 165 |
-
|
| 166 |
-
|
| 167 |
-
placeholder="Modifique el texto para mejorar la estructura",
|
| 168 |
-
height=100,
|
| 169 |
-
on_change=update_analysis
|
| 170 |
-
)
|
| 171 |
|
| 172 |
-
|
| 173 |
-
|
| 174 |
-
|
| 175 |
-
"Analizar Cambios",
|
| 176 |
-
type="primary",
|
| 177 |
-
use_container_width=True,
|
| 178 |
-
disabled=not bool(iteration_text.strip())
|
| 179 |
-
)
|
| 180 |
|
| 181 |
-
|
| 182 |
-
|
| 183 |
-
|
| 184 |
-
with st.spinner(""): # Spinner sin texto
|
| 185 |
-
doc = nlp_models[lang_code](iteration_text)
|
| 186 |
-
analysis = perform_advanced_morphosyntactic_analysis(
|
| 187 |
-
iteration_text,
|
| 188 |
-
nlp_models[lang_code]
|
| 189 |
-
)
|
| 190 |
-
|
| 191 |
-
st.session_state.morphosyntax_state.update({
|
| 192 |
-
'current_text': iteration_text,
|
| 193 |
-
'current_analysis': {'doc': doc, 'analysis': analysis},
|
| 194 |
-
'show_iteration': True,
|
| 195 |
-
'analysis_count': st.session_state.morphosyntax_state['analysis_count'] + 1
|
| 196 |
-
})
|
| 197 |
-
|
| 198 |
-
if store_student_morphosyntax_result(
|
| 199 |
-
username=st.session_state.username,
|
| 200 |
-
text=iteration_text,
|
| 201 |
-
arc_diagrams=analysis['arc_diagrams']
|
| 202 |
-
):
|
| 203 |
-
# Actualizar solo el diagrama de iteraci贸n
|
| 204 |
-
display_morphosyntax_results(
|
| 205 |
-
st.session_state.morphosyntax_state['current_analysis'],
|
| 206 |
-
lang_code,
|
| 207 |
-
morpho_t
|
| 208 |
-
)
|
| 209 |
-
|
| 210 |
-
except Exception as e:
|
| 211 |
-
logger.error(f"Error en iteraci贸n: {str(e)}")
|
| 212 |
-
st.error("Error al procesar los cambios")
|
| 213 |
|
| 214 |
except Exception as e:
|
| 215 |
-
|
| 216 |
-
|
| 217 |
-
|
| 218 |
-
|
| 219 |
-
|
| 220 |
|
| 221 |
#########################################################################3
|
|
|
|
| 222 |
def display_morphosyntax_results(result, lang_code, morpho_t):
|
| 223 |
"""
|
| 224 |
Muestra solo el an谩lisis sint谩ctico con diagramas de arco.
|
|
|
|
|
|
|
|
|
|
|
|
|
| 225 |
"""
|
| 226 |
if result is None:
|
| 227 |
-
|
| 228 |
return
|
| 229 |
|
| 230 |
-
|
| 231 |
-
|
| 232 |
-
|
| 233 |
-
|
| 234 |
-
|
| 235 |
-
|
| 236 |
-
sentences = list(doc.sents)
|
| 237 |
-
for i, sent in enumerate(sentences):
|
| 238 |
-
with st.container():
|
| 239 |
-
st.subheader(f"{morpho_t.get('sentence', 'Sentence')} {i+1}")
|
| 240 |
try:
|
|
|
|
|
|
|
|
|
|
| 241 |
html = displacy.render(sent, style="dep", options={
|
| 242 |
"distance": 100,
|
| 243 |
"arrow_spacing": 20,
|
| 244 |
"word_spacing": 30
|
| 245 |
})
|
|
|
|
| 246 |
# Ajustar dimensiones del SVG
|
| 247 |
html = html.replace('height="375"', 'height="200"')
|
| 248 |
-
html = re.sub(r'<svg[^>]*>',
|
|
|
|
|
|
|
| 249 |
html = re.sub(r'<g [^>]*transform="translate\((\d+),(\d+)\)"',
|
| 250 |
-
lambda m: f'<g transform="translate({m.group(1)},50)"',
|
|
|
|
| 251 |
|
| 252 |
-
# Envolver en
|
| 253 |
html = f'<div class="arc-diagram-container">{html}</div>'
|
|
|
|
|
|
|
| 254 |
st.write(html, unsafe_allow_html=True)
|
|
|
|
| 255 |
except Exception as e:
|
| 256 |
-
logger.error(f"Error
|
| 257 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 42 |
import re
|
| 43 |
|
| 44 |
############################################################################
|
| 45 |
+
|
| 46 |
def display_morphosyntax_interface(lang_code, nlp_models, morpho_t):
|
| 47 |
try:
|
| 48 |
+
# CSS para layout estable
|
| 49 |
st.markdown("""
|
| 50 |
<style>
|
| 51 |
.stTextArea textarea {
|
| 52 |
font-size: 1rem;
|
| 53 |
line-height: 1.5;
|
|
|
|
|
|
|
|
|
|
|
|
|
| 54 |
min-height: 100px !important;
|
| 55 |
height: 100px !important;
|
| 56 |
}
|
| 57 |
.arc-diagram-container {
|
| 58 |
width: 100%;
|
|
|
|
|
|
|
| 59 |
padding: 0.5rem;
|
|
|
|
|
|
|
| 60 |
margin: 0.5rem 0;
|
| 61 |
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 62 |
</style>
|
| 63 |
""", unsafe_allow_html=True)
|
| 64 |
|
| 65 |
+
# Estado consistente con keys 煤nicas
|
| 66 |
+
if 'morpho_state' not in st.session_state:
|
| 67 |
+
st.session_state.morpho_state = {
|
| 68 |
+
'original_text': '',
|
| 69 |
+
'original_analysis': None,
|
| 70 |
+
'iteration_text': '',
|
| 71 |
+
'iteration_analysis': None,
|
| 72 |
+
'analysis_count': 0
|
| 73 |
+
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 74 |
|
| 75 |
+
# Texto original - Key 煤nica basada en timestamp
|
| 76 |
+
text_key = f"original_text_{st.session_state.morpho_state['analysis_count']}"
|
| 77 |
+
text_input = st.text_area(
|
| 78 |
+
"",
|
| 79 |
+
value=st.session_state.morpho_state['original_text'],
|
| 80 |
+
key=text_key,
|
| 81 |
+
height=100
|
|
|
|
| 82 |
)
|
| 83 |
|
| 84 |
+
# Bot贸n analizar original
|
| 85 |
col1, col2, col3 = st.columns([2,1,2])
|
| 86 |
with col1:
|
| 87 |
+
analyze_button = st.button(
|
| 88 |
"Analizar Texto Original",
|
| 89 |
type="primary",
|
| 90 |
+
use_container_width=True
|
|
|
|
| 91 |
)
|
| 92 |
|
| 93 |
+
# Procesar texto original
|
| 94 |
+
if analyze_button and text_input.strip():
|
| 95 |
try:
|
| 96 |
+
doc = nlp_models[lang_code](text_input)
|
| 97 |
+
analysis = perform_advanced_morphosyntactic_analysis(
|
| 98 |
+
text_input,
|
| 99 |
+
nlp_models[lang_code]
|
| 100 |
+
)
|
| 101 |
+
|
| 102 |
+
# Actualizar estado
|
| 103 |
+
st.session_state.morpho_state.update({
|
| 104 |
+
'original_text': text_input,
|
| 105 |
+
'original_analysis': {'doc': doc, 'analysis': analysis},
|
| 106 |
+
'iteration_text': text_input,
|
| 107 |
+
'analysis_count': st.session_state.morpho_state['analysis_count'] + 1
|
| 108 |
+
})
|
| 109 |
+
|
| 110 |
+
# Guardar en base de datos
|
| 111 |
+
if store_student_morphosyntax_result(
|
| 112 |
+
username=st.session_state.username,
|
| 113 |
+
text=text_input,
|
| 114 |
+
arc_diagrams=analysis['arc_diagrams']
|
| 115 |
+
):
|
| 116 |
+
# Mostrar diagrama original
|
| 117 |
+
display_arc_diagram(doc, morpho_t)
|
| 118 |
+
|
| 119 |
+
st.markdown("---")
|
| 120 |
+
|
| 121 |
+
# Campo para iteraci贸n
|
| 122 |
+
iteration_key = f"iteration_{st.session_state.morpho_state['analysis_count']}"
|
| 123 |
+
iteration_text = st.text_area(
|
| 124 |
+
"",
|
| 125 |
+
value=st.session_state.morpho_state['iteration_text'],
|
| 126 |
+
key=iteration_key,
|
| 127 |
+
height=100
|
| 128 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 129 |
|
| 130 |
+
# Bot贸n analizar iteraci贸n
|
| 131 |
+
col1, col2, col3 = st.columns([2,1,2])
|
| 132 |
+
with col1:
|
| 133 |
+
iterate_button = st.button(
|
| 134 |
+
"Analizar Cambios",
|
| 135 |
+
type="primary",
|
| 136 |
+
use_container_width=True
|
| 137 |
+
)
|
| 138 |
|
| 139 |
+
# Procesar iteraci贸n
|
| 140 |
+
if iterate_button and iteration_text.strip():
|
| 141 |
+
try:
|
| 142 |
+
doc = nlp_models[lang_code](iteration_text)
|
| 143 |
+
analysis = perform_advanced_morphosyntactic_analysis(
|
| 144 |
+
iteration_text,
|
| 145 |
+
nlp_models[lang_code]
|
| 146 |
+
)
|
| 147 |
|
| 148 |
+
# Actualizar estado
|
| 149 |
+
st.session_state.morpho_state.update({
|
| 150 |
+
'iteration_text': iteration_text,
|
| 151 |
+
'iteration_analysis': {'doc': doc, 'analysis': analysis}
|
| 152 |
+
})
|
|
|
|
|
|
|
| 153 |
|
| 154 |
+
# Guardar y mostrar
|
| 155 |
+
if store_student_morphosyntax_result(
|
| 156 |
+
username=st.session_state.username,
|
| 157 |
+
text=iteration_text,
|
| 158 |
+
arc_diagrams=analysis['arc_diagrams']
|
| 159 |
+
):
|
| 160 |
+
display_arc_diagram(doc, morpho_t)
|
|
|
|
|
|
|
|
|
|
|
|
|
| 161 |
|
| 162 |
+
except Exception as e:
|
| 163 |
+
st.error("Error al procesar iteraci贸n")
|
| 164 |
+
logger.error(f"Error: {str(e)}")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 165 |
|
| 166 |
+
except Exception as e:
|
| 167 |
+
st.error("Error al procesar texto original")
|
| 168 |
+
logger.error(f"Error: {str(e)}")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 169 |
|
| 170 |
except Exception as e:
|
| 171 |
+
st.error("Error general en la interfaz")
|
| 172 |
+
logger.error(f"Error: {str(e)}")
|
|
|
|
|
|
|
|
|
|
| 173 |
|
| 174 |
#########################################################################3
|
| 175 |
+
|
| 176 |
def display_morphosyntax_results(result, lang_code, morpho_t):
|
| 177 |
"""
|
| 178 |
Muestra solo el an谩lisis sint谩ctico con diagramas de arco.
|
| 179 |
+
Args:
|
| 180 |
+
result: Resultado del an谩lisis que contiene el documento procesado
|
| 181 |
+
lang_code: C贸digo del idioma actual
|
| 182 |
+
morpho_t: Diccionario de traducciones
|
| 183 |
"""
|
| 184 |
if result is None:
|
| 185 |
+
logger.warning("No hay resultados para mostrar")
|
| 186 |
return
|
| 187 |
|
| 188 |
+
try:
|
| 189 |
+
doc = result['doc']
|
| 190 |
+
|
| 191 |
+
with st.container():
|
| 192 |
+
sentences = list(doc.sents)
|
| 193 |
+
for i, sent in enumerate(sentences):
|
|
|
|
|
|
|
|
|
|
|
|
|
| 194 |
try:
|
| 195 |
+
st.subheader(f"{morpho_t.get('sentence', 'Sentence')} {i+1}")
|
| 196 |
+
|
| 197 |
+
# Renderizar con opciones optimizadas
|
| 198 |
html = displacy.render(sent, style="dep", options={
|
| 199 |
"distance": 100,
|
| 200 |
"arrow_spacing": 20,
|
| 201 |
"word_spacing": 30
|
| 202 |
})
|
| 203 |
+
|
| 204 |
# Ajustar dimensiones del SVG
|
| 205 |
html = html.replace('height="375"', 'height="200"')
|
| 206 |
+
html = re.sub(r'<svg[^>]*>',
|
| 207 |
+
lambda m: m.group(0).replace('height="450"', 'height="300"'),
|
| 208 |
+
html)
|
| 209 |
html = re.sub(r'<g [^>]*transform="translate\((\d+),(\d+)\)"',
|
| 210 |
+
lambda m: f'<g transform="translate({m.group(1)},50)"',
|
| 211 |
+
html)
|
| 212 |
|
| 213 |
+
# Envolver en contenedor con clase CSS
|
| 214 |
html = f'<div class="arc-diagram-container">{html}</div>'
|
| 215 |
+
|
| 216 |
+
# Mostrar diagrama
|
| 217 |
st.write(html, unsafe_allow_html=True)
|
| 218 |
+
|
| 219 |
except Exception as e:
|
| 220 |
+
logger.error(f"Error al renderizar oraci贸n {i}: {str(e)}")
|
| 221 |
+
continue # Continuar con la siguiente oraci贸n si hay error
|
| 222 |
+
|
| 223 |
+
except Exception as e:
|
| 224 |
+
logger.error(f"Error general en display_morphosyntax_results: {str(e)}")
|
| 225 |
+
|