Update app.py
Browse files
app.py
CHANGED
|
@@ -97,19 +97,24 @@ mlb = joblib.load(download_file("data/mlb_597.pkl"))
|
|
| 97 |
GO = mlb.classes_
|
| 98 |
|
| 99 |
# βββββββββββββββββββ UI βββββββββββββββββββ #
|
| 100 |
-
# --- aspecto geral da pΓ‘gina
|
| 101 |
st.set_page_config(page_title="PrediΓ§Γ£o de FunΓ§Γ΅es Moleculares de ProteΓnas",
|
| 102 |
page_icon="π§¬", layout="centered")
|
| 103 |
|
| 104 |
-
# CSS: fundo branco
|
| 105 |
st.markdown(
|
| 106 |
"""
|
| 107 |
<style>
|
| 108 |
-
body, .stApp {
|
| 109 |
-
background-color: #FFFFFF !important;
|
| 110 |
-
}
|
| 111 |
.block-container { padding-top: 1.5rem; }
|
| 112 |
textarea { font-size: 0.9rem !important; }
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 113 |
</style>
|
| 114 |
""",
|
| 115 |
unsafe_allow_html=True
|
|
@@ -118,7 +123,7 @@ st.markdown(
|
|
| 118 |
# Logo (coloca logo.png na raiz do Space)
|
| 119 |
LOGO_PATH = "logo.png"
|
| 120 |
if os.path.exists(LOGO_PATH):
|
| 121 |
-
st.image(LOGO_PATH, width=180)
|
| 122 |
|
| 123 |
st.title("PrediΓ§Γ£o de FunΓ§Γ΅es Moleculares de ProteΓnas (GO:MF)")
|
| 124 |
|
|
@@ -136,51 +141,53 @@ def parse_fasta_multiple(fasta_str):
|
|
| 136 |
if not entry.strip():
|
| 137 |
continue
|
| 138 |
lines = entry.strip().splitlines()
|
| 139 |
-
if i > 0:
|
| 140 |
header = lines[0].strip()
|
| 141 |
seq = "".join(lines[1:]).replace(" ", "").upper()
|
| 142 |
-
else:
|
| 143 |
header = f"Seq_{i+1}"
|
| 144 |
seq = "".join(lines).replace(" ", "").upper()
|
| 145 |
if seq:
|
| 146 |
parsed.append((header, seq))
|
| 147 |
return parsed
|
| 148 |
|
| 149 |
-
# βββββββββββββββββββ FUNΓΓES AUXILIARES
|
| 150 |
def go_link(go_id, name=""):
|
| 151 |
-
"""
|
| 152 |
url = f"https://www.ebi.ac.uk/QuickGO/term/{go_id}"
|
| 153 |
label = f"{go_id} β {name}" if name else go_id
|
| 154 |
return f"[{label}]({url})"
|
| 155 |
|
| 156 |
-
def
|
| 157 |
-
"""
|
| 158 |
-
|
| 159 |
-
|
| 160 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 161 |
|
| 162 |
-
# βββββββββββββββββββ FUNΓΓO PRINCIPAL DE RESULTADOS βββββββββββββββββββ #
|
| 163 |
-
def mostrar(tag, y_pred):
|
| 164 |
-
"""Mostra resultados em duas colunas dentro de um expander."""
|
| 165 |
-
with st.expander(tag, expanded=True):
|
| 166 |
col1, col2 = st.columns(2)
|
| 167 |
|
| 168 |
-
#
|
| 169 |
with col1:
|
| 170 |
st.markdown(f"**GO terms com prob β₯ {THRESH}**")
|
| 171 |
hits = mlb.inverse_transform((y_pred >= THRESH).astype(int))[0]
|
| 172 |
if hits:
|
| 173 |
for go_id in hits:
|
| 174 |
name, defin = GO_INFO.get(go_id, ("β sem nome β", ""))
|
| 175 |
-
defin = re.sub(r'^\\s
|
| 176 |
defin or "")
|
| 177 |
-
st.markdown(f"- {go_link(go_id, name)}
|
| 178 |
if defin:
|
| 179 |
st.caption(defin)
|
| 180 |
else:
|
| 181 |
st.code("β nenhum β")
|
| 182 |
|
| 183 |
-
#
|
| 184 |
with col2:
|
| 185 |
st.markdown(f"**Top {TOP_N} GO terms mais provΓ‘veis**")
|
| 186 |
for rank, idx in enumerate(np.argsort(-y_pred[0])[:TOP_N], start=1):
|
|
@@ -197,23 +204,22 @@ if predict_clicked:
|
|
| 197 |
|
| 198 |
for header, seq in parsed_seqs:
|
| 199 |
with st.spinner(f"A processar {header}β¦ (pode demorar alguns minutos)"):
|
| 200 |
-
#
|
| 201 |
emb_pb = embed_seq(FINETUNED_PB, seq, CHUNK_PB)
|
| 202 |
emb_bfd = embed_seq(FINETUNED_BFD, seq, CHUNK_PB)
|
| 203 |
emb_esm = embed_seq(BASE_ESM, seq, CHUNK_ESM)
|
| 204 |
|
| 205 |
-
#
|
| 206 |
y_pb = mlp_pb.predict(emb_pb)
|
| 207 |
y_bfd = mlp_bfd.predict(emb_bfd)
|
| 208 |
y_esm = mlp_esm.predict(emb_esm)[:, :597] # alinhar nΒΊ de termos
|
| 209 |
|
| 210 |
-
# ββββββββββββ STACKING ββββββββββββ #
|
| 211 |
X = np.concatenate([y_pb, y_bfd, y_esm], axis=1)
|
| 212 |
y_ens = stacking.predict(X)
|
| 213 |
|
| 214 |
-
mostrar(
|
| 215 |
|
| 216 |
-
# βββββββββββββββββββ LISTA COMPLETA
|
| 217 |
with st.expander("Mostrar lista completa dos 597 GO terms possΓveis", expanded=False):
|
| 218 |
cols = st.columns(3)
|
| 219 |
for i, go_id in enumerate(GO):
|
|
|
|
| 97 |
GO = mlb.classes_
|
| 98 |
|
| 99 |
# βββββββββββββββββββ UI βββββββββββββββββββ #
|
|
|
|
| 100 |
st.set_page_config(page_title="PrediΓ§Γ£o de FunΓ§Γ΅es Moleculares de ProteΓnas",
|
| 101 |
page_icon="π§¬", layout="centered")
|
| 102 |
|
| 103 |
+
# CSS: fundo branco + logo + traΓ§o vertical entre colunas
|
| 104 |
st.markdown(
|
| 105 |
"""
|
| 106 |
<style>
|
| 107 |
+
body, .stApp { background-color: #FFFFFF !important; }
|
|
|
|
|
|
|
| 108 |
.block-container { padding-top: 1.5rem; }
|
| 109 |
textarea { font-size: 0.9rem !important; }
|
| 110 |
+
|
| 111 |
+
/* traΓ§o vertical: primeiro column do par recebe border-right */
|
| 112 |
+
div[data-testid="column"]:first-child {
|
| 113 |
+
border-right: 1px solid #E0E0E0;
|
| 114 |
+
padding-right: 1.2rem;
|
| 115 |
+
}
|
| 116 |
+
/* empurra logo ligeiramente para baixo */
|
| 117 |
+
img.logo { margin-top: 0.5rem; }
|
| 118 |
</style>
|
| 119 |
""",
|
| 120 |
unsafe_allow_html=True
|
|
|
|
| 123 |
# Logo (coloca logo.png na raiz do Space)
|
| 124 |
LOGO_PATH = "logo.png"
|
| 125 |
if os.path.exists(LOGO_PATH):
|
| 126 |
+
st.image(LOGO_PATH, width=180, output_format="auto", use_column_width=False, caption=None, clamp=False, channels="RGB", format="PNG", class_name="logo")
|
| 127 |
|
| 128 |
st.title("PrediΓ§Γ£o de FunΓ§Γ΅es Moleculares de ProteΓnas (GO:MF)")
|
| 129 |
|
|
|
|
| 141 |
if not entry.strip():
|
| 142 |
continue
|
| 143 |
lines = entry.strip().splitlines()
|
| 144 |
+
if i > 0:
|
| 145 |
header = lines[0].strip()
|
| 146 |
seq = "".join(lines[1:]).replace(" ", "").upper()
|
| 147 |
+
else:
|
| 148 |
header = f"Seq_{i+1}"
|
| 149 |
seq = "".join(lines).replace(" ", "").upper()
|
| 150 |
if seq:
|
| 151 |
parsed.append((header, seq))
|
| 152 |
return parsed
|
| 153 |
|
| 154 |
+
# βββββββββββββββββββ FUNΓΓES AUXILIARES βββββββββββββββββββ #
|
| 155 |
def go_link(go_id, name=""):
|
| 156 |
+
"""Markdown link para QuickGO."""
|
| 157 |
url = f"https://www.ebi.ac.uk/QuickGO/term/{go_id}"
|
| 158 |
label = f"{go_id} β {name}" if name else go_id
|
| 159 |
return f"[{label}]({url})"
|
| 160 |
|
| 161 |
+
def uni_url(header):
|
| 162 |
+
"""URL UniProt (primeiro token do header)."""
|
| 163 |
+
return f"https://www.uniprot.org/uniprotkb/{header.split()[0]}"
|
| 164 |
+
|
| 165 |
+
# βββββββββββββββββββ MOSTRAR RESULTADOS βββββββββββββββββββ #
|
| 166 |
+
def mostrar(header, y_pred):
|
| 167 |
+
"""Expander com botΓ£o UniProt + duas colunas de resultados."""
|
| 168 |
+
with st.expander(header, expanded=True):
|
| 169 |
+
# botΓ£o que abre UniProt noutra aba
|
| 170 |
+
st.link_button("π Ver na UniProt", uni_url(header), type="primary")
|
| 171 |
+
st.markdown("---")
|
| 172 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 173 |
col1, col2 = st.columns(2)
|
| 174 |
|
| 175 |
+
# --- coluna 1 : termos acima do threshold
|
| 176 |
with col1:
|
| 177 |
st.markdown(f"**GO terms com prob β₯ {THRESH}**")
|
| 178 |
hits = mlb.inverse_transform((y_pred >= THRESH).astype(int))[0]
|
| 179 |
if hits:
|
| 180 |
for go_id in hits:
|
| 181 |
name, defin = GO_INFO.get(go_id, ("β sem nome β", ""))
|
| 182 |
+
defin = re.sub(r'^\\s*\"?(.+?)\"?\\s*(\\[[^\\]]*\\])?\\s*$', r'\\1',
|
| 183 |
defin or "")
|
| 184 |
+
st.markdown(f"- {go_link(go_id, name)}")
|
| 185 |
if defin:
|
| 186 |
st.caption(defin)
|
| 187 |
else:
|
| 188 |
st.code("β nenhum β")
|
| 189 |
|
| 190 |
+
# --- coluna 2 : top-N mais provΓ‘veis
|
| 191 |
with col2:
|
| 192 |
st.markdown(f"**Top {TOP_N} GO terms mais provΓ‘veis**")
|
| 193 |
for rank, idx in enumerate(np.argsort(-y_pred[0])[:TOP_N], start=1):
|
|
|
|
| 204 |
|
| 205 |
for header, seq in parsed_seqs:
|
| 206 |
with st.spinner(f"A processar {header}β¦ (pode demorar alguns minutos)"):
|
| 207 |
+
# ---------- EMBEDDINGS ----------
|
| 208 |
emb_pb = embed_seq(FINETUNED_PB, seq, CHUNK_PB)
|
| 209 |
emb_bfd = embed_seq(FINETUNED_BFD, seq, CHUNK_PB)
|
| 210 |
emb_esm = embed_seq(BASE_ESM, seq, CHUNK_ESM)
|
| 211 |
|
| 212 |
+
# ---------- PREDIΓΓES ----------
|
| 213 |
y_pb = mlp_pb.predict(emb_pb)
|
| 214 |
y_bfd = mlp_bfd.predict(emb_bfd)
|
| 215 |
y_esm = mlp_esm.predict(emb_esm)[:, :597] # alinhar nΒΊ de termos
|
| 216 |
|
|
|
|
| 217 |
X = np.concatenate([y_pb, y_bfd, y_esm], axis=1)
|
| 218 |
y_ens = stacking.predict(X)
|
| 219 |
|
| 220 |
+
mostrar(header, y_ens)
|
| 221 |
|
| 222 |
+
# βββββββββββββββββββ LISTA COMPLETA βββββββββββββββββββ #
|
| 223 |
with st.expander("Mostrar lista completa dos 597 GO terms possΓveis", expanded=False):
|
| 224 |
cols = st.columns(3)
|
| 225 |
for i, go_id in enumerate(GO):
|