Update app.py

app.py

@@ -318,16 +318,21 @@ def find_float_after_label(text: str, label: str, max_numbers: int = 2, window:
     return out
 
 def extract_pdf_name(text: str):
-    patterns = [
-        r"Personne:\s*([A-Za-zÀ-ÿ\- ]+),\s*\d{2}/\d{2}/\d{4}",
-        r"Personne:\s*([A-Za-zÀ-ÿ\- ]+)",
-    ]
-    for pattern in patterns:
-        m = re.search(pattern, text, flags=re.S)
-        if m:
-            return m.group(1).strip()
-    return None
+    if not text:
+        return None
 
+    for line in text.splitlines():
+        clean = " ".join(line.replace("\xa0", " ").split())
+        if "Personne:" in clean:
+            after = clean.split("Personne:", 1)[1].strip()
+            # on coupe avant la date si elle existe
+            m = re.match(r"(.+?)(?:,\s*\d{2}/\d{2}/\d{4})?$", after)
+            if m:
+                name = m.group(1).strip()
+                if name:
+                    return name
+
+    return None
 
 def parse_zebris_pdf(uploaded_pdf):
     text = extract_text_from_pdf(uploaded_pdf)
@@ -358,68 +363,97 @@ def parse_zebris_pdf(uploaded_pdf):
         "fore_peak_time_pct_d": np.nan,
     }
 
-    # On normalise le texte pour éviter les problèmes de retours ligne / espaces
-    text_flat = re.sub(r"\s+", " ", text).replace("\xa0", " ").strip()
+    # normalisation forte du texte
+    text_flat = text.replace("\xa0", " ")
+    text_flat = re.sub(r"\s+", " ", text_flat).strip()
 
     num = r"(\d+,\d+|\d+\.\d+|\d+)"
 
     def to_float(x):
         return float(x.replace(",", "."))
 
-    def extract_pair(section_label, sub_label=None):
+    def extract_gd_pair(block_text, label):
         """
-        Extrait une paire Gauche / Droite du type :
-        ... Gauche Droite 1066,5±67,8 1040,5±49,3 ...
+        Cherche un pattern du type:
+        label ... Gauche Droite 1066,5±67,8 1040,5±49,3
+        et renvoie (gauche, droite)
         """
-        if sub_label:
-            pattern = (
-                rf"{re.escape(section_label)}.*?"
-                rf"{re.escape(sub_label)}\s+Gauche\s+Droite\s+"
-                rf"{num}\s*±\s*{num}\s+{num}\s*±\s*{num}"
-            )
-        else:
-            pattern = (
-                rf"{re.escape(section_label)}\s+Gauche\s+Droite\s+"
-                rf"{num}\s*±\s*{num}\s+{num}\s*±\s*{num}"
-            )
-
-        m = re.search(pattern, text_flat, flags=re.S)
+        pattern = (
+            rf"{re.escape(label)}\s*Gauche\s*Droite\s*"
+            rf"{num}\s*±\s*{num}\s*"
+            rf"{num}\s*±\s*{num}"
+        )
+        m = re.search(pattern, block_text, flags=re.S)
         if not m:
             return np.nan, np.nan
-
-        # groupes 1 et 3 = valeurs moyennes G / D
         return to_float(m.group(1)), to_float(m.group(3))
 
-    # 1) Transition (section spécifique)
-    tg, td = extract_pair("Instant du passage du talon vers l'avant-pied, s")
-    data["transition_g"], data["transition_d"] = tg, td
-
-    # 2) Forces max
-    fg, fd = extract_pair("Force maximale, N", "Forefoot (Three zones)")
-    mg, md = extract_pair("Force maximale, N", "Midfoot (Three zones)")
-    hg, hd = extract_pair("Force maximale, N", "Heel (Three zones)")
+    # on isole le bloc utile de la page 8
+    start_idx = text_flat.find("Analyse du pieds en trois zones")
+    if start_idx == -1:
+        start_idx = text_flat.find("Analyse du pieds en trois zones")
+    if start_idx == -1:
+        start_idx = text_flat.find("Modification de la charge")
 
-    data["fore_force_g"], data["fore_force_d"] = fg, fd
-    data["mid_force_g"], data["mid_force_d"] = mg, md
-    data["heel_force_g"], data["heel_force_d"] = hg, hd
+    end_idx = text_flat.find("Durée de contact au sol", start_idx if start_idx != -1 else 0)
 
-    # 3) Pressions max
-    fpg, fpd = extract_pair("Pression maximale, N/cm²", "Forefoot (Three zones)")
-    mpg, mpd = extract_pair("Pression maximale, N/cm²", "Midfoot (Three zones)")
-    hpg, hpd = extract_pair("Pression maximale, N/cm²", "Heel (Three zones)")
-
-    data["fore_pressure_g"], data["fore_pressure_d"] = fpg, fpd
-    data["mid_pressure_g"], data["mid_pressure_d"] = mpg, mpd
-    data["heel_pressure_g"], data["heel_pressure_d"] = hpg, hpd
+    if start_idx != -1 and end_idx != -1 and end_idx > start_idx:
+        zone_block = text_flat[start_idx:end_idx]
+    elif start_idx != -1:
+        zone_block = text_flat[start_idx:]
+    else:
+        zone_block = text_flat
 
-    # 4) Timing pic de force
-    ftg, ftd = extract_pair("Instant pic de force, % de phase d'appui", "Forefoot (Three zones)")
-    mtg, mtd = extract_pair("Instant pic de force, % de phase d'appui", "Midfoot (Three zones)")
-    htg, htd = extract_pair("Instant pic de force, % de phase d'appui", "Heel (Three zones)")
+    # 1. Transition
+    tg, td = extract_gd_pair(zone_block, "Instant du passage du talon vers l'avant-pied, s")
+    data["transition_g"], data["transition_d"] = tg, td
 
-    data["fore_peak_time_pct_g"], data["fore_peak_time_pct_d"] = ftg, ftd
-    data["mid_peak_time_pct_g"], data["mid_peak_time_pct_d"] = mtg, mtd
-    data["heel_peak_time_pct_g"], data["heel_peak_time_pct_d"] = htg, htd
+    # 2. Force maximale
+    fg, fd = extract_gd_pair(zone_block, "Forefoot (Three zones)")
+    # attention: Forefoot apparaît plusieurs fois dans la page, donc on fait des blocs plus précis
+    force_section_match = re.search(
+        rf"Force maximale,\s*N(.*?)Pression maximale,\s*N/cm²",
+        zone_block,
+        flags=re.S
+    )
+    if force_section_match:
+        force_section = force_section_match.group(1)
+        fg, fd = extract_gd_pair(force_section, "Forefoot (Three zones)")
+        mg, md = extract_gd_pair(force_section, "Midfoot (Three zones)")
+        hg, hd = extract_gd_pair(force_section, "Heel (Three zones)")
+        data["fore_force_g"], data["fore_force_d"] = fg, fd
+        data["mid_force_g"], data["mid_force_d"] = mg, md
+        data["heel_force_g"], data["heel_force_d"] = hg, hd
+
+    # 3. Pression maximale
+    pressure_section_match = re.search(
+        rf"Pression maximale,\s*N/cm²(.*?)Instant pic de force,\s*% de phase d'appui",
+        zone_block,
+        flags=re.S
+    )
+    if pressure_section_match:
+        pressure_section = pressure_section_match.group(1)
+        fpg, fpd = extract_gd_pair(pressure_section, "Forefoot (Three zones)")
+        mpg, mpd = extract_gd_pair(pressure_section, "Midfoot (Three zones)")
+        hpg, hpd = extract_gd_pair(pressure_section, "Heel (Three zones)")
+        data["fore_pressure_g"], data["fore_pressure_d"] = fpg, fpd
+        data["mid_pressure_g"], data["mid_pressure_d"] = mpg, mpd
+        data["heel_pressure_g"], data["heel_pressure_d"] = hpg, hpd
+
+    # 4. Timing pic de force
+    peak_section_match = re.search(
+        rf"Instant pic de force,\s*% de phase d'appui(.*)$",
+        zone_block,
+        flags=re.S
+    )
+    if peak_section_match:
+        peak_section = peak_section_match.group(1)
+        ftg, ftd = extract_gd_pair(peak_section, "Forefoot (Three zones)")
+        mtg, mtd = extract_gd_pair(peak_section, "Midfoot (Three zones)")
+        htg, htd = extract_gd_pair(peak_section, "Heel (Three zones)")
+        data["fore_peak_time_pct_g"], data["fore_peak_time_pct_d"] = ftg, ftd
+        data["mid_peak_time_pct_g"], data["mid_peak_time_pct_d"] = mtg, mtd
+        data["heel_peak_time_pct_g"], data["heel_peak_time_pct_d"] = htg, htd
 
     data["attaque_pdf"] = estimate_attack_from_pdf(data)
     return data
@@ -1208,7 +1242,7 @@ if uploaded_pdfs:
     for pdf in uploaded_pdfs:
         try:
             pdfs_data.append(parse_zebris_pdf(pdf))
-            st.write(pdfs_data[-1])
+            st.write("PDF parsé :", pdfs_data[-1])
         except Exception as e:
             st.warning(f"{pdf.name} : erreur lecture PDF ({e})")